Content uploaded by Yang Yang
Author content
All content in this area was uploaded by Yang Yang on Apr 02, 2015
Content may be subject to copyright.
1
Chinese residents’ demand for outbound travel:
Evidence from the Chinese Family Panel Studies data
Yang Yang1 and Xiwei Wu2
1School of Tourism and Hospitality Management, Temple University
2 Department of Demography, Renmin University of China
Yang Yang (E-mail: yangy@temple.edu), School of Tourism and Hospitality Management,
Temple University. Address: 1810 N. 13th Street, Speakman Hall 310, Philadelphia, PA
19122, USA. Phone: 1-215-204-8701.
Xiwei Wu, Ph.D (E-mail: xuei2003@163.com), Department of Demography, Renmin
University of China. Address: Department of Demography, Renmin University of China,
Beijing, 100872, China. Phone: 86-10-62514985. He is the corresponding author of this
paper.
Please cite as:
Yang, Y. and Wu, X. (2014). Chinese household demand for outbound travel:
Evidence from the China Family Panel Studies. Asia Pacific Journal of Tourism
Research, 19(10), 1111-1126.
2
Chinese residents’ demand for outbound travel:
Evidence from the Chinese Family Panel Studies
This paper investigates potential factors influencing Chinese residents’ demand for outbound travel. Based
on survey data from the Chinese Family Panel Studies (CFPS) project conducted in 2008, we utilize
several discrete choice models to analyze (1) factors explaining Chinese residents’ participation in
outbound travel in the last five years and (2) factors explaining various types of outbound travel. We
highlight the importance of age, hukou type, personal income, education level, domestic tourism
participation, foreign language proficiency, life satisfaction, and Internet use to explain Chinese residents’
outbound travel. We also observe regional differences and urban-rural differences by estimating the model
using different sub-samples. Finally, implications are presented concerning marketing efforts in targeting
potential Chinese outbound tourists.
Keywords: Chinese resident, outbound travel, discrete choice model, Chinese Family Panel Studies
Introduction
After the economic boom that followed the 1978 “reform and opening-up” policy and the
relaxation of political constraints, such as the liberalization of outbound travel and currency
regulations, Chinese outbound travel has become increasingly popular over the last decade.
By 2007, there were 134 countries and territories accessible to Chinese citizens for tourism
with an approved destination status (ADS) agreement. In 2007, 34.9 million citizens traveled
out of the country for private purposes compared with 1.12 million in 1992. The burgeoning
middle class has greatly contributed to the number of self-funded tourists traveling abroad,
3
and Chinese outbound tourists have become an important target demographic for overseas
tourism destinations (Xie & Li, 2009).
Along with the intensive growth of Chinese outbound travel, a large body of literature has
been devoted to understanding this market from different perspectives. The upsurge of
interest in this market has been reflected in many empirical studies on Chinese outbound
travelers. After reviewing past studies on Chinese outbound tourism, Cai, Li, and Knutson
(2008) found that most empirical papers conducted destination-specific data analysis. These
studies used the destination-specific survey data of Chinese tourists to investigate outbound
tourists’ profiles (Cai, Boger, & O'Leary, 1999; Cai, Lehto, & O'Leary, 2001; Ryan & Mo,
2002), motivations (Hua & Yoo, 2011; Johanson, 2008; Zhang & Lam, 1999), preferences
(Agrusa, Kim, & Wang, 2011; Chow & Murphy, 2008; Kim, Guo, & Agrusa, 2005), risk
perceptions (Teng, 2005), images (Li & Stepchenkova, 2012), and levels of satisfaction (Kau
& Lim, 2005) within the context of a particular outbound destination. However, only a
handful of studies have focused on demand side analysis or on investigating Chinese
residents’ outbound travel intentions and behaviors (Li, Lai, Harrill, Kline, & Wang, 2011;
Sparks & Pan, 2009; Zhang, Ma, & Qu, 2012).
To understand the outbound travel market of Chinese residents better, we employ two
discrete choice models to analyze the outbound travel demand from the 2008 Chinese Family
Panel Studies (CFPS) dataset. A binary logit model is first used to investigate factors
explaining whether a resident has traveled abroad over the past five years. At the same time,
regional differences and urban-rural differences are investigated by estimating a sub-sample
of each region and that of different hukou types, respectively. Next, based on the sample with
4
previous outbound travel experience, a multinomial logit model is established to understand
the sample’s outbound travel types.
Determinants of Chinese Outbound Travel
Age Age has been identified as an important indicator in characterizing Chinese outbound
travelers. Cai et al. (2001) found that more than half of US-bound Chinese tourists were in
the 31-50 age group. Guo, Kim, and Timothy (2007) found that Chinese residents over 50 are
less likely to travel abroad due to lower income, health issues, and their experiences during
difficult times under poor living conditions, among other reasons. However, Zhang et al.
(2012) found an opposite result from a demand-side empirical study and highlighted a large
proportion of aged outbound travelers. The motivations of Chinese outbound travelers are
associated with their age. Younger outbound tourists are more likely to be novelty/adventure
seekers (Kau & Lim, 2005; Li, 2007; Zhang & Lam, 1999), while older people tend to travel
abroad for visiting friends and relatives (VFR) (Jang, Yu, & Pearson, 2003; Kau & Lim,
2005).
Hukou type Hukou refers to the residency registration system the Chinese government has
used since 1958 to minimize rural-to-urban migration. Chinese citizens are divided into urban
and rural categories based on their hukou types. Urban residents have prior access to various
social welfare services provided by the government, whereas rural residents are expected to
be more self-reliant. In a word, the hukou system classifies Chinese citizens into two
apparently heterogeneous groups in terms of social welfare and economic opportunities. Due
5
to the existence of the hukou system and the urban-rural dichotomy in terms of social and
economic conditions, urban and rural residents share different consumption patterns (Yusuf,
Brooks, & Zhao, 2008) and tourism demand patterns (Wang, 2004). Therefore, we might
suspect that this urban-rural dichotomy will also be observed in outbound travel demand.
Income Various studies have found personal disposable income to be a crucial determinant of
international tourism demand (Song & Li, 2008), which is consistent with the fact that
outbound tourism is a “normal” commodity. Qu and Lam (1997) analyzed the aggregate
demand of Mainland Chinese tourists to Hong Kong and found that the per capita disposable
income is one of the two significant determinants. Personal income may also have a
non-linear influence on Chinese outbound travel. Cai et al. (2001) found two extremes of
income distribution of Chinese visitors to the U.S. Personal income is also likely to determine
one’s outbound travel motivation. In general, family/relaxation seekers (Kau & Lim, 2005),
adventure seekers (Mohsin, 2008), and travelers for financial incentives (Hua & Yoo, 2011)
usually report a lower personal income, whereas self-development seekers (Li, 2007) and
tourists for mass activities and high ‘risk’ features (Mohsin, 2008) tend to be wealthier.
Marital status As suggested by the family life-cycle theory, tourists with distinct marital
statuses tend to behave differently in overseas trips (Bojanic, 1992). By looking into the
profile of Chinese outbound travelers, Cai et al. (1999) and Johanson (2008) found that the
market is dominated by married tourists. Li, Wen, and Leung (2011) also argued that push
factors are more important for married Chinese women in choosing Hong Kong, while pull
6
factors are more important for single women. Moreover, unmarried Chinese outbound tourists
are found to place greater importance on the motivations of adventure activities (Mohsin,
2008), ego-enhancement, and communication opportunities (Hua & Yoo, 2011).
Education Tourist surveys in many overseas destinations have shown that Chinese outbound
travelers are always characterized by a higher level of education (Cai, et al., 1999; Cai, et al.,
2001; Ryan & Mo, 2002). In a demand side study, Zhang et al. (2012) also found that
potential outbound travelers from Shanghai are well-educated. This result can be explained
by well-educated people being more willing to explore novelty in foreign countries trigged by
their accumulated knowledge and are able to collect travel information from distinct channels.
Well-educated outbound tourists are more likely to be motivated by ego-enhancement and
communication opportunities (Hua & Yoo, 2011). Moreover, Mohsin (2008) found that
Chinese outbound tourists with higher education rated all motivation items higher, suggesting
a higher degree of involvement during outbound trips.
Foreign language proficiency Because international travel is inherently associated with
uncertainties and risks, proficiency in the local language of a destination is important in
reducing the uncertainties and risks of traveling in an unfamiliar environment. This language
barrier is especially significant for Chinese tourists (Sparks & Pan, 2009) because most
Chinese residents cannot communicate in a second language. Moreover, because the Chinese
are characterized by a high level of uncertainty avoidance (Mok & Defranco, 2000), they are
more likely to avoid uncertain situations, such as traveling to a place with significant
7
language barriers. Among 16 travel constraints listed in research by Li, Zhang, Mao, and
Deng (2011), language was perceived as the greatest barrier for Chinese residents’ outbound
travel.
Domestic tourism experience Residents’ past domestic tourism experiences can also be used
to predict their outbound travel demand. Those who have abundant experiences with
domestic travel are familiar with the decision-making processes involved, including how to
collect travel information, how to book airline tickets and hotel rooms, and how to set up a
traveling itinerary according to their preferences. Moreover, for those who favor domestic
tourism, their accumulated traveling experiences could expand their traveling scope,
triggering interests in and intentions to travel abroad.
Life satisfaction A large body of literature has confirmed the contribution of tourism to the
overall life satisfaction of tourists (Milman, 1998; Sirgy, Kruger, Lee, & Yu, 2011). It has
been suggested that people with a higher level of life satisfaction are more likely to
participate in various tourism activities. Gilbert and Abdullah (2004) found that people in the
holiday-taking group have a higher sense of well-being before the holiday than those in the
non-holiday-taking group. Outbound travel is a type of activity that can enhance people’s
sense of happiness and achievement. Therefore, people who are satisfied with and possess a
positive attitude toward life are more likely to take part in outbound travel.
Internet use The Internet plays a substantial and long-lasting role as a source of information
8
for Chinese residents’ outbound travel (Sparks & Pan, 2009), and this source is especially
important for adventure/pleasure seekers (Kau & Lim, 2005) and novelty and knowledge
seekers (Li, 2007). Potential outbound travelers can collect information in a convenient and
efficient way through tourism websites, social network sites (SNS), Internet forums, and
online chatting. These Internet users possess a great advantage when making outbound travel
decisions and, hence, are more likely to travel abroad.
Among the determinants discussed above, the influences of hukou type, foreign language
proficiency, domestic tourism experience, life satisfaction, and Internet use on Chinese
outbound travel have not been studied in any rigorous quantitative research. More
importantly, we use various micro-econometric models to investigate multiple determinants
in each model, an approach that is different from traditional methods that test factors
separately, such as analysis of variance (ANOVA), the independent sample t-test, and the
Chi-square test.
Data Source
CFPS is a systematic project managed by the Institute of Social Science Survey (ISSS),
Peking University. The CFPS gathers both economic and non-economic information on
respondents with particular emphases on economic activities, education outcomes, family
dynamics and relationships, migration, and health status. The data we used in this paper were
obtained from the CFPS conducted in 2008, which was a pilot survey before the nation-wide
survey in 2010. By using a systematic probability proportional to size sampling, the dataset
9
consists of 2375 households, including 1120 children and 6094 adults, in 24 county-/district-
level areas of Beijing, Shanghai, and Guangdong (Institute of Social Science Survey, 2009).
Three questions in the CFPS adult-questionnaire were used to understand Chinese
residents’ outbound travel: How many times have you traveled abroad? When was your last
outbound travel? Which type was it? The responders were asked to choose from several
options from the first and third questions. For the second question on the type of outbound
travel, we generated a new category of VFR from the responder’s self-reported answer. Based
on these questions in the survey, we identified two dependent variables for empirical analysis.
The variable outbound is an indicator of residents’ outbound travel in the past: outbound = 1
if the responder had traveled abroad in the last five years, and 0 otherwise. The variable
outtype denotes the type of last outbound travel: outtype = 1 for self-funded travel for tourism
purposes; outtype = 2 for non-self-funded travel, which is typically financed by work
affiliations and governments rather than the travelers themselves; outtype = 3 for education
and work; outtype = 4 for visiting friends and relatives; and outtype = 5 for other types of
outbound travel.
(Insert Table 1 here)
We limited our sample to respondents above 16 years of age with a total count of 6054
observations. Specifically, the data cover 2180 responders in Beijing, 1759 in Shanghai, and
2115 in Guangdong. Table 1 presents a summary of the variables. From the CFPS dataset, we
find that only 3.93% of the Chinese residents in the three provinces had outbound travel
10
experience in the last five years. Of these residents, nearly half were self-funded travelers
who traveled for tourism purposes. Regional differences in outbound travel across the three
provinces were tested using the Chi-square test. As shown in Table 1, the regional difference
can be observed in the variables outbound and outtype. The dataset shows that the incidence
rate of outbound travel is highest in Shanghai, at 6.03%, and there are more self-funded than
non-self-funded outbound tourists in Guangdong.
It is of particular interest to compare our result regarding the incidence rate of Chinese
outbound travel with those from previous studies because this indicator provides valuable
information for estimating market size and potential (Li, Harrill, Uysal, Burnett, & Zhan,
2010). Zhang et al. (2012) and Sparks and Pan (2009) reported this rate to be 57.9% and
67.8%, respectively. Their overestimated incidence rates are partly due to sample selection
bias toward those residents who are more likely to travel abroad. Our estimated incidence rate
of 3.93% is also lower than the results from Li et al. (2010), which are 18.1% in Beijing, 14.1%
in Shanghai, and 12.8% in Guangzhou (the capital city of Guangdong). This difference can be
explained through different sampling schemes. In their study, telephone interviews were more
likely to cover wealthier families, whereas the CFPS dataset, with a systematic probability
proportional to size sampling, was likely to cover a more representative sample of Chinese
residents.
With regard to predictors of Chinese outbound travel demand, two continuous variables,
age and inc, are used to measure the age and annual personal income of residents (in 10000
RMB Yuan), respectively. Four independent variables are coded as 0-1 dummy variables,
including urban, married, Internet, and tourexp, where a value of one indicates urban hukou
11
type, married, Internet user, and participation in domestic tourism in the last year, respectively.
Finally, there is a set of ordinal variables to measure the socio-demographic characteristics of
the responder. edu measures the education level, ranging from the “below elementary school”
level (edu = 1) to the postgraduate level (edu = 8). language represents the foreign language
proficiency, ranging from ignorance of any foreign language (language = 1) to proficient use
of at least one foreign language (language = 4). lifesat denotes the self-rated life satisfaction,
ranging from very low (lifesat = 1) to very high (lifesat = 5). Table 1 also presents the
descriptive statistics of these independent variables. Moreover, as suggested by the significant
Wald or F statistics, regional differences do exist across the three provinces in these variables.
Econometric Model
Two discrete choice models were used to investigate Chinese residents’ outbound travel
demand: the binary logit model and the multinomial logit model. These discrete choice
models are fairly popular in modeling tourists’ behavior because they are consistent with the
random utility theory (Crouch & Louviere, 2001). At the outset, a binary logit model was
applied to analyze factors explaining Chinese residents’ participation in outbound travel. The
probability of observing a responder having outbound travel experience becomes:
exp( )
Pr( 1| ) 1 exp( )
i
ii
i
outbound
x
xx
(1)
To further explain the marginal effect of estimated coefficient β, the concept of odds is
introduced to interpret the coefficient. Odds is defined as the probability ratio of observing
outbound = 1 over outbound = 0:
12
1| 0 Pr( 1| )
( ) exp( )
Pr( 0| )
ii
ii
ii
outbound
odds outbound
x
xx
x
(2)
Therefore, the term exp(βj) – 1 can be interpreted as the factor by which the odds increase
with a one unit increase in the associated explanatory variable xj.
In the second step, a multinomial logit (MNL) model was applied to investigate different
types of Chinese residents’ outbound travels. The MNL model is a well-received model for
capturing multiple nominal outcomes, and it can be thought of as incorporating several binary
logit models with all pair-wise comparisons among all outcomes (Long & Freese, 2006). The
probability of outbound travel type m is specified as:
5
1
exp( )
Pr( | ) exp( )
m i m
ii
j i j
j
outtype m
z
xz
, m = 1, …, 5 (3)
Results
Binary Logit Model for Outbound Travel
The estimation results of the binary logit model for outbound travel are presented in Table 2.
To capture the non-linear influences of some variables (Yang, Wong, & Zhang, 2011), the
squared terms of age, inc, and lifesat were included as age2, inc2, and lifesat2, respectively.
Furthermore, the dummy variables Beijing and Shanghai were incorporated to capture
regional differences in outbound travel, leaving Guangdong as the reference category.
Because there were missing values in some explanatory variables, Model 1 was fitted based
on 4843 observations from all the three provinces. Most estimated coefficients in Model 1 are
statistically significant with expected signs. The results suggest that a Chinese resident who
participated in domestic tourism in the last year, with an urban hukou type, a higher level of
13
education, a higher level of foreign language proficiency, and who use the Internet frequently
is more likely to have outbound travel experience in the past five years. The dummy variable
Beijing is estimated to be significant and negative, suggesting that, ceteris paribus, the
incidence rate of outbound travel in Guangdong (reference category) is higher than that in
Beijing. This result can be explained by the advantageous location of Guangdong, which is
close to outbound destinations such as Hong Kong and Macau.
(Insert Table 2 here)
We apply the concept of odds to interpret the estimated coefficients in the model (Equation
2). The estimated coefficient of urban is 0.720, suggesting that, other variables being constant,
the odds of past outbound travel for urban residents is 105.44% higher than the odds for rural
residents. In the same way, the estimated coefficient of Internet, 0.511, indicates that, after
controlling for other factors, the odds of past outbound travel for Internet users is 66.70%
higher than the odds for non-Internet users. Moreover, the estimated coefficient of language,
0.276, indicates that a one-level increase in foreign language proficiency contributes to a
31.78% increase in the odds of past outbound travel, which highlights the great importance of
foreign language proficiency in determining Chinese residents’ outbound travel.
To elaborate the estimated coefficients of the three squared terms, age2, inc2, and lifesat2,
we graphed the predicted probabilities of outbound travel (incidence rate) over a range of
values after setting all other variables at the mean value. Figure 1 demonstrates the non-linear
effect of age on the incidence rate of outbound travel. The “all-sample” curve shows a
14
positive relationship between age and outbound travel for all Chinese residents before the age
of 70. Figure 2 shows the non-linear effect of personal income. The corresponding
“all-sample” curve is consistently upward, suggesting the positive effect of personal income
on the incidence rate of outbound travel. Finally, Figure 3 shows the non-linear effect of life
satisfaction, which indicates a turning point in the "all-sample" curve at lifesat = 4. This result
suggests that when residents' life satisfaction is not very high, the incidence rate of outbound
travel increases with life satisfaction. However, for those resident who are very satisfied with
life (lifesat = 5), their average incidence rate is actually lower than those with a moderately
high level of life satisfaction (lifesat = 4).
(Insert Figure 1 here)
(Insert Figure 2 here)
(Insert Figure 3 here)
We also estimated the model with a sub-sample of each province to understand regional
differences in outbound travel. In Table 2, Models 2, 3, and 4 were estimated based on
observations in Beijing, Shanghai, and Guangdong, respectively. The variable married is still
not significant in any model, suggesting that marital status is not a significant determinant of
Chinese residents’ outbound travel. It is important to highlight the significance of urban, edu,
tourexp, language, and Internet in the three models. The results indicate that well-educated
residents and those with a higher level of foreign language proficiency in Beijing, Internet
users in Shanghai, and urban residents in Guangdong are far more likely to have outbound
15
travel experience.
To understand the nonlinear influences of age, personal income, and life satisfaction, we
also used the probability curves to illustrate their impacts. Figure 1 highlights the different
nonlinear impacts of age across the three provinces. The Shanghai curve moves consistently
upward with age, highlighting a constant positive influence of age on the incidence rate of
outbound travel for Shanghai residents. The curves of Beijing and Guangdong exhibit an
“inverse-U” shape. Residents who are older than 65 are less likely to travel abroad. This
result can be explained by the decline in income and the deterioration of health of aged
residents. In Figure 2, the steeper curve for Shanghai suggests that personal income exerts a
more substantial marginal influence on outbound travel for the residents. In Figure 3, the
curves of the different provinces imply that the marginal influence of life satisfaction is
greatest in Shanghai, and residents with a moderately high level of life satisfaction (lifesat =
4) are most likely to have outbound travel experience. However, the curve of Guangdong is
relatively flat, illustrating a moderate impact of life satisfaction.
For the different hukou types, Models 5 and 6 in Table 2 fitted the urban and rural resident
samples, respectively. By comparing the estimates of these models, we observe substantial
differences between the outbound travel demand of urban and rural residents. For example,
the coefficients of edu, tourexp, and Internet are estimated to be statistically significant in
Model 5 but are insignificant in Model 6. Moreover, by comparing the slopes of the various
probability curves in Figures 1 to 3, we find that the marginal impacts of age and personal
income are greater for urban residents, while the marginal impact of life satisfaction is fairly
similar between urban and rural residents.
16
Multinomial Logit Model for Outbound Travel Types
The results from the MNL model are presented in Table 3. Because this model only covers
residents with outbound travel experience in the last five years, it has a small sample size,
with only 203 observations in the three provinces. To make the statistical inference reliable
for a small sample, some insignificant explanatory variables were excluded after several
preliminary trials. Based on the estimates from this model, the social-demographic
characteristics of residents with different outbound travel types are compared. In the model,
we chose the “self-funded outbound travel for tourism purposes” (outtype = 1) type as the
baseline category. In the “outtype = 2” column of Table 3, residents with previous
“non-self-funded outbound travel” are characterized by a younger age and a lower level of
foreign language proficiency than the baseline category, as indicated by the significant and
negative coefficients for age and language. Moreover, the positive and significant
coefficients for Beijing and Shanghai imply that there are more non-self-funded outbound
tourists in Beijing and Shanghai.
For residents with previous outbound travel experience for education and work (outtype =
3), the negative and significant coefficient of age indicates that they are generally younger
than the baseline category. Moreover, residents with VFR purposes (outtype = 4) are found to
be older and less wealthy than the baseline category. This result suggests that aged people are
more inclined to travel abroad for VFR, which is consistent with previous findings (Jang, et
al., 2003; Kau & Lim, 2005). Because VFR travelers are more likely to be “inactive”
travelers (Moscardo, Pearce, Morrison, Green, & O’Leary, 2000), and their travel expenses
17
are partly covered by their friends and/or relatives at the destination, personal income is an
important consideration for this type of outbound travel.
(Insert Table 3 here)
Conclusions
In this paper, several discrete choice models were utilized to analyze the determinants of
Chinese residents’ outbound travel. The findings of this study indicate that age, hukou type,
personal income, education level, domestic tourism participation, foreign language
proficiency, life satisfaction, and Internet use are influential factors for outbound travel.
Moreover, regional differences were observed across Beijing, Shanghai, and Guangdong. For
example, education level and Internet use were observed to be important determinants of
outbound travel in Beijing and Shanghai, while hukou type played a more important role in
Guangdong. Substantial differences between urban and rural residents were also identified.
For example, age was a crucial factor for explaining the incident rate of outbound travel for
urban residents but not for rural residents.
This paper’s results provide an extremely clear picture of the socio-demographic
characteristics of Chinese outbound travelers and have implications for several important
marketing strategies for Chinese outbound destinations. First, because language was
highlighted as a key barrier to outbound travel, it is necessary for outbound destinations to
provide more tourism materials and guidance in Chinese during the trip. Second, the results
underscored the higher outbound travel propensity of Internet users and well-educated
18
residents. Therefore, online marketing is important for improving the share of this major
segment. Third, because regional differences were observed in Chinese outbound travel,
different marketing strategies should be proposed for different regions. In Beijing, for
example, destination marketing organizations should devote a higher level of marketing
attention to well-educated residents, while in Guangdong they should target urban residents.
Finally, because self-funded outbound travelers for tourism purposes were generally more
involved in domestic tourism, as well, outbound destinations can cooperate with domestic
destinations to strengthen their image for potential travelers to attract this particular segment.
This paper contributed to the current body of literature on Chinese outbound tourism in
several ways. First, we used large-scale household survey data, for the first time in the
literature, to further understanding of Chinese outbound travel from the demand side. The
dataset we utilized provided comprehensive socio-demographic information for thousands of
Chinese households and reduced the sample selection bias toward potential outbound
travelers in past surveys. Second, we set up several discrete choice models to compare the
profile difference between residents with and without outbound traveling experience as well
as to analyze the type of residents’ past outbound travels. Therefore, the results substantially
contribute to the evaluation of potential opportunities for outbound tourism. Finally, to the
best of our knowledge, this study represents one of the first efforts at highlighting the
regional differences in Chinese outbound travel by exploring these differences in three major
areas.
Because the CFPS data are not designed specifically for tourism research, there is a limited
set of information regarding the level of respondents’ tourism experience. More importantly,
19
we cannot obtain detailed trip-related characteristics, which might be another set of
determinants influencing outbound travel patterns. In addition, the sample analyzed was not a
nationally representative sample. Our results might reflect only a part of China, specifically
the three most developed areas. Therefore, we recommend that future studies apply the model
to fit the nation-wide outbound travel demand data with more detailed trip-related
information.
References
Agrusa, J., Kim, S. S., & Wang, K.-C. (2011). Mainland Chinese tourists to Hawaii: Their
characteristics and preferences. Journal of Travel & Tourism Marketing, 28(3),
261-278.
Bojanic, D. C. (1992). A look at modernized family life cycle and overseas travel. Journal of
Travel & Tourism Marketing, 1(1), 61-79.
Cai, L. A., Boger, C., & O'Leary, J. (1999). The Chinese travelers to Singapore, Malaysia,
and Thailand: A unique Chinese outbound market. Asia Pacific Journal of Tourism
Research, 3(2), 2-13.
Cai, L. A., Lehto, X. Y., & O'Leary, J. (2001). Profiling the U.S.-bound Chinese travelers by
purpose of trip. Journal of Hospitality & Leisure Marketing, 7(4), 3-16.
Cai, L. A., Li, M., & Knutson, B. J. (2008). Research on China outbound market: A
meta-review. Journal of Hospitality & Leisure Marketing, 16(1-2), 5-20.
Chow, I., & Murphy, P. (2008). Travel activity preferences of Chinese outbound tourists for
overseas destinations. Journal of Hospitality & Leisure Marketing, 16(1-2), 61-80.
Crouch, G. I., & Louviere, J. J. (2001). A review of choice modeling research in tourism,
20
hospitality and leisure. In J. A. Mazanec, G. I. Crouch, J. R. B. Ritchie & A. G.
Woodside (Eds.), Consumer psychology of tourism, hospitality and leisure (Vol. 2, pp.
67-86). Wallingford, UK: CABI Publishing.
Gilbert, D., & Abdullah, J. (2004). Holidaytaking and the sense of well-being. Annals of
Tourism Research, 31(1), 103-121.
Guo, Y., Kim, S. S., & Timothy, D. J. (2007). Development characteristics and implications
of Mainland Chinese outbound tourism. Asia Pacific Journal of Tourism Research,
12(4), 313-332.
Hua, Y., & Yoo, J. J.-E. (2011). Travel motivations of Mainland Chinese travelers to the
United States. Journal of China Tourism Research, 7(4), 355-376.
Institute of Social Science Survey. (2009). Chinese Family Dynamics 2009. Beijing: Peking
University Press.
Jang, S., Yu, L., & Pearson, T. (2003). Chinese travellers to the United States: A comparison
of business travel and visiting friends and relatives. Tourism Geographies, 5(1),
87-108.
Johanson, M. M. (2008). The outbound Mainland China market to the United States:
Uncovering motivations for future travel to Hawaii. Journal of Hospitality & Leisure
Marketing, 16(1-2), 41-59.
Kau, A. K., & Lim, P. S. (2005). Clustering of Chinese tourists to Singapore: An analysis of
their motivations, values and satisfaction. International Journal of Tourism Research,
7(4-5), 231-248.
Kim, S. S., Guo, Y., & Agrusa, J. (2005). Preference and positioning analyses of overseas
21
destinations by Mainland Chinese outbound pleasure tourists. Journal of Travel
Research, 44(2), 212-220.
Li, M. (2007). Modeling the travel motivation of Mainland Chinese outbound tourists.
Unpublished Ph.D Dissertation, Purdue University, West Lafayette, IN.
Li, M., Wen, T., & Leung, A. (2011). An exploratory study of the travel motivation of
Chinese female outbound tourists. Journal of China Tourism Research, 7(4), 411-424.
Li, M., Zhang, H., Mao, I., & Deng, C. (2011). Segmenting Chinese outbound tourists by
perceived constraints. Journal of Travel & Tourism Marketing, 28(6), 629-643.
Li, X., Harrill, R., Uysal, M., Burnett, T., & Zhan, X. (2010). Estimating the size of the
Chinese outbound travel market: A demand-side approach. Tourism Management,
31(2), 250-259.
Li, X., Lai, C., Harrill, R., Kline, S., & Wang, L. (2011). When east meets west: An
exploratory study on Chinese outbound tourists’ travel expectations. Tourism
Management, 32(4), 741-749.
Li, X., & Stepchenkova, S. (2012). Chinese outbound tourists’ destination image of America.
Journal of Travel Research, 51(3), 250-266.
Long, J. S., & Freese, J. (2006). Regression models for categorical dependent variables using
Stata (2nd ed.). College Station, TX: Stata Press.
Milman, A. (1998). The impact of tourism and travel experience on senior travelers'
psychological well-being. Journal of Travel Research, 37(2), 166-170.
Mohsin, A. (2008). Analysis of Chinese travellers' attitudes toward holidaying in New
Zealand: The impact of socio-demographic variables. Journal of Hospitality &
22
Leisure Marketing, 16(1-2), 21-40.
Mok, C., & Defranco, A. L. (2000). Chinese cultural values: Their implications for travel and
tourism marketing. Journal of Travel & Tourism Marketing, 8(2), 99-114.
Moscardo, G., Pearce, P., Morrison, A., Green, D., & O’Leary, J. T. (2000). Developing a
typology for understanding visiting friends and relatives markets. Journal of Travel
Research, 38(3), 251-259.
Qu, H., & Lam, S. (1997). A travel demand model for Mainland Chinese tourists to Hong
Kong. Tourism Management, 18(8), 593-597.
Ryan, C., & Mo, X. (2002). Chinese visitors to New Zealand—Demographics and
perceptions. Journal of Vacation Marketing, 8(1), 13-27.
Sirgy, M. J., Kruger, P. S., Lee, D.-J., & Yu, G. B. (2011). How does a travel trip affect
tourists’ life satisfaction? Journal of Travel Research, 50(3), 261-275.
Song, H., & Li, G. (2008). Tourism demand modelling and forecasting-A review of recent
research. Tourism Management, 29(2), 203-220.
Sparks, B., & Pan, G. W. (2009). Chinese outbound tourists: Understanding their attitudes,
constraints and use of information sources. Tourism Management, 30(4), 483-494.
Teng, W. (2005). Risks perceived by Mainland Chinese tourists towards Southeast Asia
destinations: A fuzzy logic model. Asia Pacific Journal of Tourism Research, 10(1),
97-115.
Wang, N. (2004). The rise of touristic consumerism in urban China. Tourism Recreation
Research, 29(2), 47-58.
Xie, Y., & Li, M. (2009). Development of China's outbound tourism and the characteristics of
23
its tourist flow. Journal of China Tourism Research, 5(3), 226-242.
Yang, Y., Wong, K. K. F., & Zhang, J. (2011). Determinants of length of stay for domestic
tourists: Case study of Yixing. Asia Pacific Journal of Tourism Research, 16(6),
619-633.
Yusuf, F., Brooks, G., & Zhao, P. (2008). Household Consumption in China: An Examination
of the Utility of Urban-Rural Segmentation. In S. H. Murdock & D. A. Swanson
(Eds.), Applied Demography in the 21st Century (pp. 285-298). Netherlands: Springer
Zhang, H. Q., & Lam, T. (1999). An analysis of Mainland Chinese visitors’ motivations to
visit Hong Kong. Tourism Management, 20(5), 587-594.
Zhang, Y., Ma, E., & Qu, H. (2012). An explorative analysis of Shanghai residents' outbound
travel preferences during global financial crisis. Anatolia, 1-13.
24
Table 1. Summary of Variables from the CFPS Dataset
Variable
All-sample
Beijing
Shanghai
Guangdong
Wald or F statistic
(df)
outbound = 1
3.93%
3.07%
6.03%
3.07%
28.81(2)***
outtype = 1
46.54%
31.65%
41.92%
68.75%
42.94(8)***
outtype = 2
32.26%
39.57%
36.53%
18.75%
outtype = 3
13.82%
20.14%
14.37%
6.25%
outtype = 4
5.76%
5.76%
5.99%
5.47%
outtype = 5
1.61%
2.88%
1.20%
0.78%
age
46.38
46.76
48.56
44.16
33.91(2,6022)***
inc
19.84
16.76
26.80
16.42
8.38(2,4976)***
urban
54.42%
60.82%
71.40%
33.40%
606.71(2)***
married
84.11%
83.10%
87.09%
82.67%
16.65(2)***
Internet
29.64%
29.06%
34.68%
26.02%
34.99(2)***
tourexp
32.71%
35.36%
39.09%
24.75%
99.20(2)***
edu
3.54
3.93
3.65
3.04
156.67(2,6016)***
language
1.87
1.91
2.00
1.72
45.32(2,6034)***
lifesat
3.43
3.72
3.32
3.23
147.89(2,6088)***
Sample size
6054
2180
1759
2115
(Notes: *** indicates p<0.01)
25
Table 2. Estimation Results of Binary Logit Models
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
age
0.103***
0.176**
0.0707
0.135*
0.108***
0.0738
(0.0352)
(0.0768)
(0.0494)
(0.0769)
(0.0366)
(0.139)
age2
-0.000708*
*
-0.00141*
-0.000362
-0.00104
-0.000734*
*
-0.000494
(0.000334)
(0.000733)
(0.000457)
(0.000772)
(0.000348)
(0.00132)
urban
0.720**
-0.477
0.559
1.402***
(0.287)
(0.622)
(0.468)
(0.440)
married
-0.162
-0.336
-0.287
-0.124
-0.263
0.903
(0.317)
(0.652)
(0.450)
(0.657)
(0.330)
(1.299)
inc
0.143***
0.210**
0.168***
0.163***
0.126***
0.184*
(0.0349)
(0.101)
(0.0506)
(0.0548)
(0.0374)
(0.105)
inc2
-0.00172*
-0.00652
-0.00261***
-0.00169**
-0.00153
-0.000586
(0.000947)
(0.00399)
(0.000938)
(0.000779)
(0.00100)
(0.00313)
edu
0.311***
0.732***
0.355***
-0.00803
0.321***
0.331
(0.0655)
(0.150)
(0.0924)
(0.138)
(0.0658)
(0.295)
tourexp
1.059***
0.556
1.243***
1.565***
1.124***
0.552
(0.202)
(0.368)
(0.310)
(0.443)
(0.216)
(0.737)
language
0.276**
0.565***
-0.00707
0.361*
0.300***
0.0287
(0.111)
(0.210)
(0.169)
(0.213)
(0.115)
(0.386)
lifesat
1.169**
1.003
1.661*
0.900
0.979*
5.427**
(0.527)
(1.041)
(0.872)
(0.904)
(0.525)
(2.540)
lifesat2
-0.162**
-0.163
-0.209*
-0.142
-0.139*
-0.733**
(0.0740)
(0.145)
(0.121)
(0.136)
(0.0742)
(0.366)
Internet
0.511**
0.391
0.883***
-0.000663
0.475**
1.027
(0.209)
(0.451)
(0.304)
(0.415)
(0.211)
(0.959)
Beijing
-0.470**
-0.539**
-0.0973
(0.223)
(0.235)
(0.690)
Shanghai
0.0157
-0.0399
0.111
(0.210)
(0.224)
(0.615)
constant
-11.71***
-14.89***
-11.81***
-11.16***
-10.71***
-19.30***
(1.261)
(2.455)
(2.142)
(2.335)
(1.315)
(5.824)
Sample
size
4843
1542
1561
1740
2695
2148
Sample
All
Beijing
Shanghai
Guangdon
g
Urban-huk
ou
Rural-huk
ou
Pseudo R2
0.247
0.310
0.243
0.259
0.193
0.161
AIC
1279.796
353.432
551.540
377.632
1102.986
194.538
BIC
1377.075
422.863
621.130
448.633
1185.574
273.950
(Notes: * indicates p<0.10, ** indicates p<0.05, *** indicates p<0.01. Robust standard errors are
presented in parentheses.)
26
Table 3. Estimation Results of Multinomial Logit Model
outtype = 2
outtype = 3
outtype = 4
outtype = 5
age
-0.0493***
-0.0355*
0.0637**
0.0685
(0.0151)
(0.0204)
(0.0308)
(0.0631)
edu
0.228
0.0246
0.00580
0.0318
(0.141)
(0.219)
(0.317)
(0.266)
married
0.573
0.331
-1.384
9.446***
(0.640)
(0.692)
(1.238)
(1.526)
inc
0.0277
0.0426
-0.344*
-1.471*
(0.0245)
(0.0278)
(0.203)
(0.755)
tourexp
-0.182
-0.431
-0.933
15.02***
(0.450)
(0.647)
(0.740)
(0.902)
language
-0.424*
0.278
-0.138
-0.0337
(0.244)
(0.300)
(0.413)
(0.452)
Beijing
2.014***
2.371***
1.984*
-13.16***
(0.601)
(0.819)
(1.190)
(1.430)
Shanghai
1.732***
2.069***
1.882*
-0.0578
(0.523)
(0.773)
(1.048)
(1.694)
constant
-0.441
-2.731
-4.065
-29.39***
(1.109)
(1.907)
(2.907)
(2.787)
Sample size
203
Pseudo R2
0.172
AIC
462.655
BIC
581.931
(Notes: * indicates p<0.10, ** indicates p<0.05, *** indicates p<0.01. Robust standard errors are
presented in parenthesis. outtype = 1 is chosen as the baseline category, and all its coefficients are set to
0.)
