International Journal of Hospitality Management 49 (2015) 73–82

Contents lists available at ScienceDirect

International Journal of Hospitality Management

journal homepage: www.elsevier.com/locate/ijhosman

Gender wage inequality and labour mobility in the hospitality sector

Juan Antonio Campos-Soria a,∗ , Alejandro García-Pozo b , José Luis Sánchez-Ollero c
a
Department of Applied Economics, Room 1405, University of Malaga, C/ El Ejido, No. 6, CP: 29013, Málaga, Spain
b
Department of Applied Economics, Room 1410, University of Malaga, C/ El Ejido, No. 6, CP: 29013, Málaga, Spain
c
Department of Applied Economics, Room 1407, University of Malaga, C/ El Ejido, No. 6, CP: 29013, Málaga, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The main aim of this paper is twofold. Firstly we analyse the effects of educational mismatch on workers’
Received 17 November 2014 occupational mobility and secondly, we quantify the contribution of educational mismatch and labour
Received in revised form 6 April 2015 mobility on gender wage inequality. The decomposition of the gender wage gap was controlled by differ-
Accepted 22 May 2015
ent types of gender segregation. We use matched employer-employee data from 302 hotels in Andalusia.
The evidence suggests that not only is external mobility far higher than internal mobility in this sector,
Keywords:
but also that it is the main cause of wage inequality. The reason for this is the existence of labour dis-
Educational mismatch
crimination against women, which manifests as a lack of access to labour improvements under the same
Labour mobility
Horizontal segregation
conditions as men. Educational mismatch has a limited effect on internal and external mobility. Finally,
Vertical segregation gender discrimination is the overall cause of gender wage inequality and is mainly due to horizontal
Gender wage gap segregation and external mobility.
Hospitality © 2015 Elsevier Ltd. All rights reserved.

1. Introduction
The importance of the tourism sector to the Spanish economy is
undeniable. Spain began to increase its popularity as a tourism des-
tination in the 1960s and since then, with few exceptions, tourism
has positively contributed to the national economy by levelling the
balance of payments and becoming a source of employment. The
tourism sector accounted for 10.8% of the Spanish GDP in 2011 (INE,
2013) and 11.8% of total employment in Spain (IET, 2013). Accord-
ing to data from the Bank of Spain, the tourism surplus for 2012
had a coverage rate of 122.5% for the trade deficit and 280.5% for
the current account deficit, while the commercial coverage rate of
tourism was the highest this century. In 2013, the World Trade
organisation ranked Spain second in revenues from international
tourism, just behind the United States, and third in international
arrivals, after France and the United States, and before China (WTO,
2013). In the case of Andalusia, the contribution of tourism to GDP
was estimated at 12.8% in 2013. In this year, there was an annual
average of 329,600 individuals working in this sector, representing
an increase of +2.4% compared to 2012. These workers represent
13.0% of the 2.54 million employed in the region and 16.5% of those
∗ Corresponding author. Fax: +34 952136616.
E-mail addresses: jacampos@uma.es (J.A. Campos-Soria), ag@uma.es
(A. García-Pozo), jlsanchez@uma.es (J.L. Sánchez-Ollero).
employed in the Andalusian service sector. These data show the
great importance of the tourism industry to employment.
These facts contrast with the specific features of most tourism
activities, which are characterized by a large number of unskilled
jobs, high staff turnover, few prospects for promotion, relatively
low wages, precarious job security, and seasonal employment. The
phenomenon of temporary contracts mainly affects women, young
people, and low-skilled workers (Jimeno, 2005). In the hotel indus-
try, four out of ten women and three out of ten men have temporary
contracts (Malo and Muñoz-Bullón, 2008). As in the case of workers
with temporary contracts, the most vulnerable workers with fewer
choices dominate the group with part-time contracts (Pedraza et al.,
2010). The reasons underlying this situation can be found in the
social behaviour of workers and in the interests of the companies in
this sector. From the company’s point of view, the key term is usu-
ally “labour flexibility”, whereas from the workers’ point of view the
key term is usually “conciliation”. Labour flexibility is understood
as the need for part-time labour to meet peak demand during work-
ing hours (such lunch-time and supper-time in the restaurant, or
cleaning and tidying the customers’ rooms). The combination of the
interests of both parties has led to the empirical results obtained by
various authors: women can obtain part-time jobs that allow them
to balance such work with family needs (Rosendaal, 2003).
Thus, on the one hand, the Spanish tourism sector is very impor-
tant due to its great capacity to generate employment even in
times of economic crisis. On the other hand, much of the tourism
workforce is female, includes occupations that are almost exclu-

http://dx.doi.org/10.1016/j.ijhm.2015.05.009
0278-4319/© 2015 Elsevier Ltd. All rights reserved.
74 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82

sively held by women, and is characterised by low labour mobility Table 1

Characteristics of employment in the Andalusian hospitality industry by gender.
(Marchante et al., 2007), a high degree of occupational segregation
(Campos-Soria et al., 2011), and a high degree of gender wage dis- Andalusia
crimination (Campos-Soria et al., 2009; García-Pozo et al., 2014).
2006 2010
According to García-Pozo et al. (2014), the impact of gender on
wages is very important in the hospitality industry, as shown by Variables Male Female Male Female

the fact that there is a 13.38% wage penalty for female workers Age
compared to male workers. These authors have suggested that this 16–24 years old 19.78 27.18 7.94 9.73
25–54 years old 73.99 71.14 81.59 8525
wage penalty may be mainly caused by the type of jobs performed
More than 54 years old 6.23 1.68 10.47 5.01
by women. Most of these jobs involve subordinate tasks, excluding
technical or management tasks. Most of the wage discrimination is Contracts
Permanent 75.09 65.77 81.95 82.01
caused by the occupational segregation of women into jobs such as
Temporary 24.91 34.23 18.05 17.99
waitressing or cleaning, which are at the lowest levels of responsi-
Contracts
bility.
Full-time 83.15 58.72 87.36 81.12
Despite the importance of the tourism sector and the role of Part-time 16.85 41.28 12.64 18.88
women within it, few studies have taken into account the gender
Employer size
perspective and are limited because of the lack of data that would
Fewer than 20 workers 23.81 15.44 23.47 30.09
allow a complete analysis of this specific sector. Thus, there is a From 20 to 99 workers 53.11 55.7 20.22 24.48
clear need for an in-depth analysis of employment in the Spanish More than 99 workers 23.08 28.86 56.32 45.43
tourism sector and the role of women within it. In contrast to other Mean years of schooling of workers 7.61 7.65 8.06 8.16
studies, we met this need by using a database containing a large Nominal wage (D /hour) 9.51 7.89 10.69 8.47
Real wage (at 2006 prices) 10.24 8.31 9.82 7.79
representative set of data on employees containing disaggregated
information related to jobs rather than occupations. Source: INE and own calculations.

This article analyses from a gender perspective two fundamental

human capital aspects in the Spanish hospitality industry: Firstly,
we investigate the reasons for the discrepancy between actual and
required levels of education, and the resulting differences in returns
to education, using a human capital mobility framework; Secondly,
based on an explicitly theoretical approach, we decompose the
gender wage gap and draw attention to the effects of educational
mismatch, occupational segregation (distinguishing between hor-
izontal and vertical segregation), and internal and external labour
mobility on the gender wage difference. As far as we know, this is
the first study to jointly assess the relative impact of educational
mismatch, labour mobility, and horizontal and vertical segregation
on the gender wage difference in the hospitality industry.
This article is organized as follows: the next section draws the
context of our research; after that, we present the literature review;
the theoretical model and empirical specification is developed in
Section 4; a case study is presented in Section 5; the main results are
described in Section 6; and some concluding remarks are presented
in the final section.
2. The context
The economic crisis is having a devastating effect on the Span-
ish economy in general and a significant impact on the hospitality
sector in Andalusia in particular. According to the latest annual
Andalusian Tourism Balance Sheet, which is published by the
regional government, from the time the economic crisis began job
losses have been more pronounced among female workers in this
sector than among the male workers for three consecutive years
(2009–2011). Table 1 shows data obtained from the Wage Structure
Survey for 2006 and 2010, which are the latest available for this
quadrennial database compiled by the Spanish National Institute
of Statistics (INE). The table shows the evolution of some variables
that can be used to assess the impact on labour mobility and on
the gender pay gap in the Andalusian hotel sector of moving from a
situation of strong economic growth (the Andalusian GDP grew by
4.2% in 2006) to one of a reduction in regional production (in 2010,
this growth became negative at −1.7% according to the INE).
The first important finding shown in Table 1 is the increase
in the average age of Andalusian workers, given that the eco-
nomic crisis has significantly reduced youth employment, whereas
there has been an increase in the participation of workers 25 years
old or more. This finding is consistent with the increase in youth
unemployment rate in Andalusia, which increased from 20.79% in
2006 to 49.57% in 2010 (INE). In recent years, the participation
of women in the hospitality sector has increased. Statistics show
that, after controlling for other factors, countries with a higher
share of part-time contracts tend to have higher female partici-
pation. Although the Spanish government has attempted to reduce
the incidence of temporary employment by reducing payroll taxes
and dismissal costs for permanent contracts, the disparity in the
Spanish labour market persists, with women being more affected
in Spain than women in other countries. However, in the hospi-
tality industry, the crisis has affected the behaviour of managers
when hiring workers. On the one hand, permanent contracts have
increased compared to temporary contracts, especially in the case
of women, and secondly, part-time contracts have increased com-
pared to full-time contracts. Again, the latter situation has had a
far greater impact on women. These changes may be due to the
elimination from the Andalusian hospitality industry of precarious
employment, which is less expensive to eliminate than full-time
and permanent employment. Medium-sized companies have been
more affected by this type of job destruction than small and large
businesses. The latter types of company may have characteristics
that make them more adaptable to market circumstances.
Regarding the educational level of employees, it is noteworthy
that women in Andalusian hospitality industry have an increased
average number of years of formal education and that during the
crisis this difference has increased. These two facts have helped
to reduce the gender pay gap (Marchante et al., 2005; Campos-
Soria et al., 2009). The data shown in Table 1 suggest the existence
of a gender pay gap in the Andalusian hospitality industry. With
the advent of the economic crisis, this gender difference has even
increased in real terms from a differential of 23.23% in 2006–26.06%
in 2010.
3. Literature review
In contrast to other sectors, the hospitality industry is char-
acterized by a higher percentage of female employees and
an occupational structure with lower educational attainment
(Marchante el al., 2005). The literature is unanimous in showing
that women workers receive lower wages, although this is linked
to the fact that they work in sectors and jobs with lower basic
 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82
wages (Lee and Kang, 1998). The gender wage gap in the Span-
ish hospitality industry has been demonstrated by Campos-Soria
et al. (2009) and García-Pozo et al. (2012), among others. The fac-
tors that underlie the wage gap include the propensity of women to
work in part-time jobs, occupational segregation, and educational
and work experience mismatches.
Many studies have analysed the determinants of educational
mismatch and the relationships between educational mismatch,
labour mobility, and wages in the workforce market. The existence
of permanent educational mismatch suggests that the resources
invested in education have been wrongly allocated, which could
penalize the productivity of the workforce. Therefore, it is impor-
tant to identify whether any mismatch is permanent or temporary,
to analyse how mismatch is related to labour mobility, and to
empirically verify its effect on wages. The literature shows that
overeducated and undereducated workers earn lower and higher
wages, respectively, than adequately educated workers (i.e. work-
ers whose level of education is appropriate to the job). There is
also empirical evidence showing the relationship between educa-
tional mismatch and gender discrimination, with women suffering
greater wage penalties that are associated with educational mis-
match (Salinas-Jiménez et al., 2014).
Regarding labour mobility, the Career Mobility Theory (Rosen,
1972; Sicherman and Galor, 1990) suggests that workers might be
interested in temporarily taking a job for which they are overqual-
ified (an entry job) in order to acquire enough experience to move
to better paid jobs or jobs with a greater chance of promotion.
In this sense, any mismatch would be temporary because overe-
ducation would encourage employee mobility. The Job Matching
Theory (Jovanovic, 1979; Sicherman, 1991) views overeducation as
being disadvantageous to employees. That is, overqualified work-
ers would try to find a better match by changing their job within
the firm or by moving to another firm. Consequently, this theory
predicts greater voluntary mobility among overqualified workers
and thus overeducation is considered a temporary mismatch. In
contrast, undereducation can be considered an advantageous mis-
match because it promotes permanence in the same job and within
the same firm.
According to García-Pozo et al. (2011), the main reason workers
change firms in the hospitality sector is to improve their work-
ing conditions (61% of total quits), rather than to improve their
upward occupational mobility (23% of total quits). Marchante et al.
(2007) pointed out that Spanish hospitality workers are mainly
characterized by interfirm turnover (81%, voluntary plus invol-
untary turnover), rather than by intrafirm turnover (31%). These
figures are in line with the predictions of the skill model of the
weak internal labour market presented by Riley (1997). However,
male and female workers present different patterns, which may
reinforce horizontal and vertical segregation (Jordan, 1997). Men
focus on earning higher wages more than women and use external
mobility to achieve this aim, whereas women mainly focus on their
labour conditions in the same firm (Marchante et al., 2007).
There is broad empirical evidence that gender segregation has
an impact on the gender wage gap (Bayard et al., 2003 or Gupta
and Rothstein, 2005; among others). Gender segregation can be
addressed at the industrial, establishment, and occupational level.
In each case, segregation refers to the over-representation of
women in given sectors, establishments, or occupations. In the case
of tourism, most studies use dummy variables to explain the effects
of these types of female segregation on wage differentials (Delfim
and Varejao, 2007; Thrane, 2008) and thus it does not allow to iden-
tify which part of wage inequality is due to different kinds of gender
segregation. As far as we know, Campos-Soria et al. (2009) are the
only authors to have investigated the effect of different types of
gender segregation, including horizontal and vertical segregation,
on the gender wage gap in the hospitality industry.
75
Labour conditions considerably differ between men and women
in terms of wages, educational level, labour mobility, and gender
segregation in the hospitality sector compared to the economy as a
whole. Thus, this article attempts to provide insight, from a gender
perspective, into the relationships between schooling mismatch,
workers’ mobility across firms and occupations in the hospitality
sector, and wage patterns.
4. Methodology
4.1. Model specification
The econometric model used to address the external and inter-
nal mobility framework is an adaptation of the occupational
mobility approach proposed by Sicherman (1991). The econometric
specification is as follows:

Yi∗ = xki ˇk + i (1)
This equation is based on a random utility model, where Yi∗ is the
latent variable of either internal or external mobility, which depend
on exogenous variables xki , ˇk is a vector of unknown parameters,
and i is the error term. These equations are estimated using a
binary logit model (BLM), which is a suitable model since for each
equation (external and internal mobility), we can define a depen-
dent variable Yi , in binary terms. That is, the dependent variable for
internal labour mobility takes value 1 if the worker has changed job
within the current firm, and 0 otherwise, whereas the dependent
variable for the external labour mobility takes value 1 if the worker
has moved to another firm and 0 otherwise.
In our case, the logit model uses a logistic distribution function
such that the specification of these models is as follows:
1 e˛+ˇk xki
Yi = + i = + i (2)
1 + e−˛−ˇk xki 1 + e˛+ˇk xki
The logit models assume a nonlinear relationship between the
explanatory variables and the probability of a given event, since
increases in the probability are not always equal given that these
increases depend on the original level of the independent variables.
The linearization of the model makes the estimated parameters
easier to interpret. Thus, based on Eq. (2), the probability of internal
or external mobility, denoted by Li , can be obtained as follows:
e˛+ˇk xki
E(Yi ) = Prob (Yi = 1) = Li = (3)
1 + e˛+ˇk xki
From Eq. (3), it is easy to obtain the odds ratio, which is the ratio
between the probability of an event occurring versus the probabil-
ity of it not occurring:
Li
OR = = e˛+ˇk xki (4)
(1 − Li )
Regarding the specification of the gender wage gap decomposi-
tion, the theoretical literature refers to discrimination as a wage gap
between men and women that is not due to differences in produc-
tivity. In order to appropriately specify the relationship between
wages and different observable characteristics, wage regressions
have to be estimated separately for men and women. Eqs. (5) and
(6) show these estimates:

ln (wmi ) = xmi ˇm + mi (5)
  
ln wfi = xfi ˇf + fi (6)
where xmi and xfi are the vectors of men and women’s charac-
teristics, respectively, ˇm and ˇf are the coefficient vectors to be
estimated for each group, and mi and fi are the error terms for men
76 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82
and women, respectively. Under the assumption that the observ-
able characteristics of each gender yield different returns, Oaxaca
and Ransom (1994) decomposed wages by gender as follows:
     
     
ln wm − ln wf = xm − xf ˇ∗ + xm ˇ
ˆ m − ˇ∗ − xf
 
where ln (wm ) and ln wf are the geometric means of the log-
arithm of the net wage per effective working hour for men and
 
women, respectively, xm and xf are vectors for the geometric means
ˆ m and ˇ
ˆ f are
of the observable characteristics for each gender, and ˇ
the estimated coefficient vectors for the male and female samples
separately. ˇ∗ is the coefficient vector of the wage structure in the
absence of discrimination. We use Neumark (1988) proposal, which
assumes the estimates obtained for the whole sample as the wage
structure in non-discriminatory markets. The decomposition pro-
posed in Eq. (7) enables the wage gap to be broken down in two
parts. The first part captures the proportion of the gender pay gap
that is due to differences in productivity. The second part shows the
contribution of gender discrimination that is due to wage advan-
tages for men and wage disadvantages for women in relation to the
wage structure in the absence of discrimination.
4.2. Description of the variables
The exogenous variables considered in the mobility and wage
regressions can be grouped into personal and human charac-
teristics and firm and job characteristics. The inclusion of these
explanatory variables related to the individuals, their positions, or
the establishments is justified by the fact that they allow worker
productivity to be controlled. Becker (1964) human capital the-
ory and the wage equation proposed in this context by Mincer
(1974) consider that wage differences are caused by differences
in the workers’ marginal productivity. These authors include indi-
vidual control variables that explain these disparities. Subsequent
empirical studies, such as those by Bayard et al. (2003), Gupta and
Rothstein (2005), and Simón et al. (2008), have included additional
control variables related to the worker’s position and the establish-
ment’s characteristics. Personal characteristics include age, gender,
and marital status. The latter two variables are dummies that take
value 1 if the employee is male and married, respectively, and 0
otherwise. Human capital variables include educational attainment
expressed as the number of schooling years and dummy variables
indicating vocational training in tourism, tenure in the firm, and
tenure in the job. These variables measure real experience rather
than potential experience since women have a higher probability
of leaving the labour market.
The set of firm and job control variables considered in the equa-
tions are the size of the establishment, type of contract, horizontal
and vertical segregation, level of responsibility in the job, the for-
mal education of the worker in relation to the job requirements,
and the internal and external mobility of the workers. A dummy
variable is used to classify establishments as small (less than 25
employees), medium (25–100 employees), or large (more than 100
employees). The variable full-time permanent contract takes value
1 when the worker has a full-time permanent contract and 0 oth-
erwise. In this way, we attempted to assess whether this type of
contract entails higher wages than temporary or part-time con-
tracts. In Spain, both types of contract are used in the hospitality
industry according to the establishment adjusting labour demand
to fluctuations in tourism demand.
The construction of the horizontal and vertical segregation vari-
ables requires grouping jobs into different functional areas and
responsibility levels. We consider six functional areas (reception,
administration, kitchen, catering, cleaning, and maintenance) and
five levels of responsibility1 (from level 4, which includes jobs
with the least responsibility, to level 0, which denotes manage-
rial positions). The horizontal segregation variable assigns to each

individual the percentage of women working in each functional
area within each level of responsibility in the hotels or restau-
ˆ f − ˇ∗
ˇ (7)
rants with the same category.2 The vertical segregation variable
assigns to each individual the percentage of women working in
each level of responsibility within each functional area in the
same category. Educational mismatch is calculated using an indi-
rect subjective method. This method collects the employees’ and
employers’ opinions regarding the educational level needed to per-
form the job. Thus, the comparison between needed and attained
education yields three worker categories: adequately educated
refers to workers with the correct level of attained education
needed to perform the job; undereducated refers to workers whose
level of attained education is lower than that needed to perform the
job; and overeducated refers to workers whose level of attained
education is higher than that needed to perform the job.
5. Case study
Data were obtained from a research project that was developed
in 2010 by an interdisciplinary team from the University of Malaga
in collaboration with a group of establishments located in Andalusia
(southern Spain). The database included representative parame-
ters from 2476 employees working in 302 hotels (rated as 3, 4,
and 5 stars) with more than seven workers. In total, there were
784 hotels in these categories in the region, and thus the sample
formed 38.5% of the total number of establishments. The informa-
tion was obtained from semi-structured personal questionnaires
administered to the managers and hospitality workers. The aim of
the survey was to generate a representative sample of the hospital-
ity sector in Andalusia. Table 2 shows the most relevant personal
and human capital characteristics and firm and job characteristics
by type of labour mobility and gender.
Strikingly, external mobility (81.32%) was far higher than inter-
nal mobility (30.47%) in the whole sample. This difference is
understandable given that this sector is characterized by a very
low internal mobility and high external mobility (Burns, 1993;
Jameson, 2000). García-Pozo et al. (2011) have suggested that
although skills are transferable between firms, there are barriers to
changing occupations within hotels and catering establishments.
For example, a waiter cannot simply become a receptionist and
receptionist cannot become a cook. Occupational rigidity affects
all workers regardless of their educational level, and thus work-
ers acquire skills through deliberate mobility between firms. This
implies a weak internal labour market that allows requirements
and labour turnover fluctuations to match supply and demand
(Riley and Szivas, 2003).
On the other hand, the percentage distribution of overedu-
cated (17.92%), undereducated (27.61%), and adequately educated
(54.47%) workers in the sample was consistent with percentages
distributions obtained in other studies of the Spanish hospital-
ity sector (García-Pozo et al., 2014). Educational mismatch had a
greater impact on internal mobility than on external mobility, both
for men (51.7% versus 49.22%) and women (48.06% versus 40.98%).
Internal and external labour mobility reached their highest val-
ues in medium companies and among workers with medium-high
levels of responsibility. However, in the latter case, these values
1
Jobs are classified into functional areas and levels of responsibility following the
Nationwide Labor Agreement for the Hospitality Sector and the Provincial Collective
Agreement for the Hospitality Sector in;1; Spain.
2
Note that if these percentages are calculated for the whole sample, the resulting
horizontal segregation variable includes part of the vertical segregation, as levels of
responsibility are not homogeneous in all the areas.
 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82 77

Table 2
Average values of the variables used by type of labour mobility and gender.

Internal mobility External mobility Whole sample

Variables Men Women Men Women

Personal and human capital characteristics

Gendera 67.59 32.41 65.42 34.58 61.33
Marital statusa 77.19 56.38 69.36 51.75 59.68
Schooling yearsb 9.28 10.07 9.19 9.47 9.32
Vocational training in tourisma 40.21 37.97 34.86 27.62 31.86
Tenure in the firmb 13.96 9.97 7.97 5.19 7.84
Tenure in current jobb 9.40 6.74 6.93 4.66 6.71
Ageb 39.06 36.31 36.94 34.03 35.77

Firm and job characteristics

Small firma 17.64 18.69 27.38 28.09 26.51
Medium firma 61.00 62.28 55.10 54.77 55.98
Large firma 21.36 19.03 17.52 17.14 17.51
Responsibility level 0a 4.95 3.91 3.41 1.47 2.67
Responsibility level 1a 30.65 14.33 19.96 6.12 13.71
Responsibility level 2a 54.95 47.23 58.34 40.51 51.75
Responsibility level 3a 7.12 30.29 11.98 41.13 23.44
Responsibility level 4a 2.32 4.24 6.31 10.77 8.43
Full-time and permanent contracta 72.42 61.41 55.40 39.81 50.48
Adequately educateda 48.30 52.94 50.78 59.02 54.47
Under-educateda 42.13 26.47 35.02 19.38 27.61
Over-educateda 9.57 21.59 14.20 21.60 17.92
Internal mobilitya – – – – 30.47
External mobilitya – – – – 81.32
a
Expressed as a percentage.
b
Years.

were higher among men than women. In addition, full-time per-
manent contracts favoured internal mobility more than external
labour mobility in both genders.
Regarding personal and human capital characteristics, the aver-
age years of schooling for both men and women were higher for
internally mobile workers; however, compared to men, a higher
level of formal education (more than 10 schooling years) had been
attained by women who were internally or externally mobile.
Finally, in relation to both types of mobility, the mean values of
tenure in the firm and in the job were much higher among men
than women.
6. Results
6.1. Labour mobility estimations
Table 3 shows the results of estimating the econometric speci-
fications presented above for internal and external labour mobility
by gender. A binomial logit model was used based on the method
of maximum likelihood. In line with McFadden (1974), we present
the estimated coefficients, odds-ratios and marginal effects for
coefficients that were statistically significant. Goodness-of-fit was
evaluated using McFadden’s (1974) pseudo-R2 , which showed
acceptable values in all cases. The values of the Likelihood Ratio
(LR) Chi-Square test and its associated probability suggest that
the estimated coefficients of the explanatory variables have sig-
nificant effects on the probability of the dependent variable. The
observed values of the dependent variable were correctly classi-
fied by the models to match its predicted value (higher than 77% in
all the estimated equations). The area under the Receiver Operat-
ing Characteristic (ROC) curve in all the logit estimations presented
exceeded 0.79. These data confirm that the models had good pre-
dictive capability. The reference variables for each set of dummy
variables created from a categorical variable are as follows: small
firm, responsibility level 0 (maximum level), and adequately edu-
cated workers. The probability of either gender being internally
mobile was influenced by practically the same explanatory vari-
ables. Although in some cases there were quantitative differences,
they are always in the same direction. All other factors remaining
equal, vocational training in tourism almost doubled the probabil-
ity of internal labour mobility among women compared to men
(11.63% versus 6.55%). The effect of an additional year of tenure
in the firm increased the probability of internal mobility, although
this effect was higher among women (14.86%) than among men
(10.09%). However, additional years in the firm decreased the prob-
ability of internal mobility, the effect again being higher in absolute
terms among women (−14.07%) than among men (−8.61%). The
signs of the estimates for both variables related to tenure matched
the predicted signs regarding internal mobility theory. Age had a
higher marginal effect on the probability of internal mobility among
men (0.92%) than among women (0.36%). It is important to note
that the last three variables analysed are continuous and thus are
of great significance regarding their predictive capability.
Compared to working in a small company, the marginal effect
of working in large establishments on the probability of internal
mobility among men was 8.99%. The marginal effect of working in
a medium company was similar for each gender (8.08%). The prob-
ability of internal mobility among workers of either gender was
significantly reduced by having lower- to middle levels of responsi-
bility compared to workers in management posts with the highest
level of responsibility. The values of the marginal effect of these
variables had a negative effect on the probability of internal labour
mobility (−33.52% for men and −36.93% for women). This could be
due to fewer years of formal education among these workers and
the type of work they perform, which reduce the opportunities for
occupational promotion in the firm. Finally, as expected, job sta-
bility increased the probability of internal mobility, although there
was a very high quantitative difference between men (8.38%) and
women (1.76%). Results again show greater difficulties for women
promotion within the company in relation to men.
The effect of statistically significant variables on the probability
of external mobility was quantitatively smaller than on the prob-
ability of internal mobility. On the other hand, the personal and
human capital characteristics of men almost exclusively determine
the probability of external mobility except in the case of having
a permanent contract, which reduces this probability by 2.64%
 78
Table 3
The determinants of internal and external labour mobility by gender.

J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82

Internal mobility External mobility

Men (Eq. (1)) Women (Eq. (2)) Men (Eq. (3)) Women (Eq. (4))

Variables ␤i Odds-ratio Marginal effect (%) ␤i Odds-ratio Marginal effect (%) ␤i Odds-ratio Marginal effect (%) ␤i Odds-ratio Marginal effect (%)

Personal and human capital characteristics

Marital status 0.0379 0.1745 0.7736c 2.1676 6.32 0.1189
Schooling years 0.0082 0.0415 −0.0925c 0.9117 −0.67 −0.0313
Vocational training in tourism 0.2863b 1.3315 6.55 0.5816c 1.7889 11.63 0.4500b 1.5684 3.07 −0.0367
Tenure in the firm 0.4466c 1.5629 10.09 0.7837c 2.1897 14.86 −0.1981c 0.8203 −1.43 −0.1403c 0.8691 −2.35
Tenure in current job −0.3808c 0.6833 −8.61 −0.7417c 0.4763 −14.07 −0.0073 −0.0252
Age 0.0405c 1.0413 0.92 0.0188c 1.019 0.36 0.1640c 1.1783 1.18 0.1025c 1.108 1.72
END>[5pt] Firm and job characteristics
Medium firm 0.2678 0.4327b 1.5415 8.08 −0.2747 −0.344
Large firm 0.3850a 1.4696 8.99 0.1823 −0.1395 0.1366
Responsibility level 1 −0.1369 −0.6234 0.0765 0.0705
Responsibility level 2 −0.8494b 0.4277 −19.32 −1.4621c 0.2317 −25.66 −0.1692 0.5052
Responsibility level 3 −1.3894c 0.2492 −25.07 −2.1887c 0.1121 −36.93 −0.2024 0.3036
Responsibility level 4 −2.4871c 0.0832 −33.52 −2.7740c 0.0624 −28.36 −0.6 −0.2043
Full-time and permanent contract 0.3757b 1.4561 8.38 0.0927b 1.0972 1.76 −0.3756a 0.6869 −2.64 0.0948
Under-educated 0.172 0.5420b 1.7194 11.05 −0.2109 −0.3008
Over-educated −0.3598 0.1674 0.0842 −0.1595
Constant −0.0892 −1.9541 −0.9727 −0.9612
McFadden’s R2 0.3073 0.3467 0.2005 0.1147
2
LR Chi (15) 617.43c 350.74c 253.27c 108.08c
Correctly classified (%) 83.8 86.27 86.29 77.22
Area under ROC curve 0.8441 0.8692 0.8174 0.7954
Observations 1562 881 1517 843
a
Level of significance 10%.
b
Level of significance 5%.
c
Level of significance 1%.
 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82 79

Table 4 19.3% in the United States. Nevertheless, these estimations are sen-
Wage regressions by gender.
sitive to the dataset and the control characteristics used. Regarding
Variables Men Women Whole sample personal and human capital variables, being married had a positive
Constant 1.5517c 1.5833c 1.5090c effect on wages for the whole sample; however, when disaggre-
Personal and human capital characteristics gated by gender the effect on wages remained positive for men but
Gender – – 0.0773c not for women. A similar effect was found for vocational training
Marital status 0.1031c -0.0068 0.0597c in tourism. Returns to formal education were quite similar for men
Schooling years 0.0156c 0.0161c 0.0168c
and women (1.56 and 1.61, respectively). A previous study using
Vocational training in tourism 0.0284b -0.0060 0.0200b
Tenure in the firm 0.0086c 0.0082c 0.0088c the OLS approach found that these returns were very low (from 2%
to 4.5%) in the hospitality sector (Marchante et al., 2005). It can be
Firm and job characteristics
Medium firm −0.0030 0.0115 0.0018a
accounted for by low educational attainment among salaried work-
Large firm 0.0952c 0.0818c 0.0920c ers in the hospitality sector relative to the entire economy3 and
Full-time and permanent contract 0.0610c 0.0561c 0.0625c because job characteristics variables were used in the specification,
Horizontal segregation 0.0016c 0.0010c 0.0013c as in previous studies. Estimated education returns in these kinds of
Vertical segregation −0.0020c −0.0016c −0.0019c
studies represent the direct effect of education after discarding the
Under-educated 0.0343c 0.0023 0.0283c
Over-educated −0.0490c −0.0651c −0.0575c effects derived from the job characteristics. Introducing these addi-
Internal mobility 0.0093 0.0429b 0.0189a tional variables reduces educational returns because they capture
External mobility 0.0721c 0.0251 0.0534c the indirect effect of education on wages. The returns estimated
R-squared 0.4012 0.3193 0.3818
for tenure in the firm were 0.86% and 0.82% for men and women,
N◦ observations 1,433 849 2,282
respectively. These percentages are much lower than those in other
All regressions are estimated by OLS. Standard errors are adjusted by heteroskedas- private service sectors, which is also in line with the results in the
ticity.
a
empirical literature (García-Pozo et al., 2014). Regarding job and
Level of significance 10%.
b
Level of significance 5%. firm characteristics, workers with a full-time permanent contract
c
Level of significance 1%. gain higher wages (a 5.61% increase among men and 6.1% among
women). The estimated coefficient for the size of the firm is statis-
tically significant and positive in all cases for establishments with
according to Eq. (3). Married workers were more likely to be exter- more than 100 employees, although it is higher among men than
nally mobile (by 6.32%) than unmarried workers. In our model, women (9.52% and 8.18%, respectively). Horizontal and vertical seg-
this variable had the greatest effect on the probability of exter- regation have an opposite effect on wages in all regressions. These
nal mobility. Another relevant finding was that a 1-year increase results are in line with the literature (Campos-Soria et al., 2009).
in schooling reduced the probability of external mobility by 0.67%. Wages are increased by female segregation in functional areas com-
This could be explained by the higher proportion in the sample of manding higher salaries, although the increase is higher for men
workers with low formal education. In relation to this result, work- than for women. However, wages are decreased by female segre-
ers with vocational training in tourism had an increased probability gation into the worst paid jobs with lower responsibility levels. This
of external mobility (3.07%) due to the demand from companies for effect is also greater in absolute terms for men than for women. In
this characteristic given that few workers have this type of train- all cases, the estimated coefficient for educational mismatch has the
ing (see Table 2). The latter three variables had no influence on the expected sign and is only statistically non-significant in the case of
probability of external mobility for women. undereducated female workers. Overeducation involves a strong
In contrast to its effect on internal mobility, an additional year wage penalty, which is greater for women than for men. Undere-
of tenure in the firm reduces the probability of external mobil- ducated male workers earn 3.43% more than adequately educated
ity among men (1.43%) and among women (2.35%), although this male workers, whereas this effect is not statistically significant for
marginal effect was much lower than in the case of internal mobil- women. Finally, the wage premium in the hospitality sector for
ity. Furthermore, compared to women, an additional year of age external and internal mobility was significant for the whole sam-
among men was associated with a positive marginal effect on exter- ple. However, there were marked differences between men and
nal mobility (1.72%). women. The literature suggests that internal mobility is low in
the hotel sector (Riley, 1997) because of the barriers to changing
6.2. Wage estimations occupations within the establishment and especially when these
occupations belong to different functional areas. Nonetheless, the
Table 4 shows the results of the estimation of the wage equa- results show that female workers are more likely than men to
tions proposed in the methodology. All the coefficients obtained are use internal mobility to increase their salaries. Women prefer to
robust regarding heteroskedasticity using iterative standard errors remain in the same company which, among those who do so, is
from White (1980). The significant difference between men and probably due to them having families. In fact, external mobility
women in human capital returns justifies the need for estimating among women does not have a positive effect on rewards. These
the wage equations under the assumption of unequal returns. The results are in line with those obtained by Marchante et al. (2007)
wage difference between men and women with the same observ- for the Spanish hospitality sector. They found that the most impor-
able characteristics and the same job is 7.73%. The wage advantage tant reason workers changed firms was to improve their working
for men is in line with the results obtained by Campos-Soria et al. conditions (61% of total quits) rather than to raise their wages (23%
(2009) for the hospitality sector in Andalusia (range 7.9%–11.1%). of the total quits). However, male co-workers were more likely to
These results are also in line with the available international liter- acquire skills through deliberate external mobility (implying higher
ature on tourism. For example, Marchante et al. (2005) estimated
the wage advantage to be 10.5% and Delfim and Varejao (2007)
estimated it to be 8.4%. However, as pointed out by Delfim and
Varejao (2007), gender discrimination in the tourism sector is lower 3
According to the estimates obtained by the Active Population Survey in Spain,
than in economy as a whole. In fact, Amuedo-Dorantes and De la the average percentage of women with schooling beyond compulsory education
Rica (2006) found that gender discrimination was 11.8% in Spain, who were employed in the hospitality sector in the period 2005–2007 was 35.92%,
whereas Bayard et al. (2003) found that it ranged from 16.2% to but 57.73% in the whole economy (34.03% and 43.18, respectively, for men).
80 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82

Table 5
Gender wage gap decomposition.

Factors Model A Model B Model C

Charact. Returns Total Charact. Returns Total Charact. Returns Total

Constant 0.00 10.65 10.65 0.00 4.92 4.92 0.00 −14.36 −14.36
Personal and human capital characteristics 56.18 33.17 89.35 52.60 25.76 78.36 48.07 29.27 77.35
Gender 39.78 – 39.78 39.97 – 39.97 35.09 – 35.09
Marital status 5.94 32.56 38.51 4.71 28.64 33.36 4.19 26.46 30.65
Schooling years −1.45 −3.00 −4.45 −1.50 −7.02 −8.52 −1.59 −1.93 −3.52
Vocational training in tourism 0.36 2.73 3.10 0.37 3.71 4.08 0.44 3.92 4.36
Tenure in the firm 11.54 0.88 12.42 9.05 0.43 9.47 9.94 0.82 10.76

Firm and job characteristics – – – 6.37 10.35 16.71 9.24 27.78 37.01
Medium sized – – – 0.00 0.00 0.00 0.00 0.00 0.00
Large sized – – – 0.00 1.82 1.82 0.00 1.08 1.08
Full-time and permanent contract – – – 4.04 −0.26 3.78 3.91 0.82 4.73
Horizontal segregation – – – −10.50 12.71 2.21 −10.36 11.91 1.55
Vertical segregation – – – 8.88 −7.26 1.62 8.83 −10.31 −1.49
Infra-educated – – – 2.03 3.09 5.13 1.90 3.48 5.38
Over-educated – – – 1.91 0.24 2.15 1.84 1.33 3.17
Internal mobility – – – – – – 0.61 −5.84 −5.23
External mobility – – – – – – 2.50 25.32 27.82
Total 56.18 43.82 100 58.96 41.04 100 57.31 42.69 100

The gender wage gap, as measured by the sex difference in mean log wages, is 0.2495 in model A, 0.2254 in model B and 0.2201 in model C. In the three models, mean
comparison tests assuming independent samples and unequal variances are applied. The numbers of each cell, represent the relative contribution of each variable obtained
from the various specifications. The cumulative relative effects of personal and capital characteristics and firm and job characteristics are shown in bold. The total relative
effects of productivity differences and return differences appear in the last line.

wages), whereas internal mobility did not have a positive effect on
their wages.
6.3. Gender wage gap decomposition
Eq. (7) was used to decompose the gender wage gap. This
decomposition was based on Neumark (1988) proposal, which
assumes that the non-discriminatory wage structure is derived
from the regression for the whole sample (i.e., the coefficients
shown in the third column, Table 4). Tables 5 shows the contri-
butions of models A–C, assuming that the returns of each variable
differ between men and women. The contributions appearing in
model C are based on the estimates for men and women shown
in Table 4. The decompositions shown in model A and B were
calculated by adding different groups of variables at successive
stages to verify the robustness of the results. Model A shows the
relative impact of each group of variables on the gender wage
gap when controlling only for personal and human capital char-
acteristics. Model B shows the estimated wage differentials by sex
when including establishment and job characteristics, but exclud-
ing mobility variables.
For each of the models, total contribution to the wage dif-
ferentials is decomposed into relative impacts of differences in
characteristics between both genders and differences in returns.
The variable under study increases the wage differential when the
value in a cell is positive and decreases it when the sign is nega-
tive. Table 5 shows that the results for all the proposals are similar,
which indicates the robustness of the empirical evidence regard-
ing gender discrimination and the relative contribution of most
of the control variables. The relative contribution of differences
in characteristics is quite similar (56.18%–58.96%). Likewise, slight
differences appear to gender discrimination. Gender discrimina-
tion under equal returns ranges from 35.09% in model C to 39.78%
in model A. If we add differences in the returns of observable char-
acteristics, the total gender discrimination under unequal returns
ranges from 77.78% in model C to 83.6% in model A, which again
demonstrates the robustness of the results. Thus, most of the wage
gap in the three models is due to gender discrimination. The latter
values are similar to those reported by Campos-Soria et al. (2009),
who found gender discrimination rates ranging from 76.61% to
77.83% in the hospitality sector.
In order to simplify the description, from this point on we will
only focus on the results from model C. Marital status increases
the wage gap by 4.19%, since being married increases the hourly
wage for men alone and there were more married men than women
in the sample. The opposite is the case regarding schooling years;
female workers have higher educational levels, which reduces the
wage difference by 1.59%. Regardless of gender, tenure in the firm
has the greatest effect on wage inequality. In fact, greater tenure
in the firm increases the wage gap by 9.94%. Earnings were also
affected by the type of contract. Workers with part-time or tem-
porary contracts earned significantly lower wages than employees
with full-time or permanent contracts. The proportion of women
with lower-paying contracts was higher than that of men, which
explains 3.91% of the wage gap. Most of the differences in job and
firm characteristics were due to horizontal and vertical segregation.
Horizontal and vertical segregation did not contribute in the same
way to increasing the wage difference, in contrast to the findings
of other descriptive studies. Female segregation in functional areas
commanding higher salaries reduces the wage gap by 10.36%. How-
ever, the segregation of women into lower levels of responsibility
contributes to the wage differential by more than 8.8%. On the other
hand, educational mismatch promotes wage inequality between
the genders. Menı́s deficit in human capital increases the wage dif-
ferential by 1.9%, whereas overeducation, which is more common
in women, raises the gap by 1.84%, since overeducation has a neg-
ative effect on wages. Finally, men are more likely to change firms
to improve their salary, which is responsible for 2.5% of the wage
gap.
The difference in returns by groups of variables reveals some
findings of interest. Marriage increases men’s wages more than
those of women, which explains most of the positive contribution
of personal and human capital characteristics (26.46%). Schooling
years reduces the wage gap by 1.93%, due to women’s greater edu-
cational return. This result is comparable to those of Thrane (2008),
who estimated that educational returns were higher for women
than for men in the tourism sector. It is also noteworthy that all
the firm and job characteristics had a positive effect on wage gap,
except for vertical segregation and internal mobility. The differ-
ence in returns from vertical segregation had the greatest effect
on reducing wage differences. Thus, differences in wage reduc-
tions caused by working in the worst paid levels of responsibility
 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82
contributed to reducing the wage differential by 10.31%. However,
horizontal segregation contributes to increasing the wage gap by
11.91%, since changing to a better paid functional area has a greater
effect on men’s wages than on women’s wages. These results are
in line with the literature that suggests that the positive contribu-
tion of occupational segregation can conceal the negative impacts
of horizontal and vertical segregation (Campos-Soria et al., 2009;
Hakim, 1992; Delfim and Varejao, 2007). Returns to educational
mismatch also raise the wage differential, since overeducation
reduces women’s wages more than men’s wages, and undered-
ucation raises men’s wages more than women’s wages. Finally,
attention should be drawn to the different returns to labour mobil-
ity variables. Since internal labour mobility only affects women’s
wages, and men are more likely to only use external mobility to
increase their wages, different returns to these variables decrease
and increase the wage differential by 5.84% and 25.32%, respec-
tively. Nevertheless, these results not only show that external
mobility is far more frequent than internal mobility in the hospi-
tality industry, but also that the external mobility has the greatest
effect on wage inequality in relation to firm and job characteristics.
7. Conclusions
The aim of this paper is twofold. Firstly, we analysed from a
gender perspective, the effects of educational mismatch on work-
ers’ occupational mobility in the hospitality industry, and secondly,
we assessed the relative impact of educational mismatch, labour
mobility, and horizontal and vertical segregation on gender wage
difference. As far as we know, this is the first study to address
both issues at the same time. This article focuses on the hos-
pitality industry, in which labour conditions differ considerably
from those in the economy as a whole. In particular, there are
marked differences in this sector between men and women in terms
of precariousness, labour mobility, occupational segregation, and
wages. Internal and external mobility were analysed using binary
choice models, and the gender wage gap was decomposed using the
Oaxaca and Ransom (1994) approach, assuming different returns to
the observable characteristics. Gender wage disparities and labour
mobility were analysed in depth by the use of establishment-
worker paired data that included information on establishments,
workers, and the job they perform thus providing a high degree
of occupational disaggregation. The results may be of relevance to
policymakers.
The descriptive data show that external mobility is much higher
than internal mobility in the hospitality industry in Andalusia. This
may be explained by the high turnover in jobs. The factors that
determine the probability of internal mobility among men and
women are mainly related to the level of responsibility held by
workers, since workers with a medium or low level of responsibil-
ity are less likely to move jobs in the current firm. Meanwhile, the
personal characteristics of individuals, particularly those related
to tenure in the firm, tenure in the current job, and vocational
training in tourism, have a lower effect among men than among
women. This may be explained by labour discrimination against
women, which manifests as a lack of access to labour improvements
under the same conditions as men. In addition, real educational
mismatch has a limited effect on the probability of changing jobs
within the firm. This result is in line with previous literature, which
found no evidence to suggest that overeducated workers obtain
better jobs in other establishments or that either gender benefits
from promotion ladders within internal labour markets (Marchante
et al., 2007). Thus, entry positions do not serve as first step in
the workers’ future career within this sector. In the present study,
undereducated women alone had an increased probability of inter-
nal mobility than adequately educated women. However, due to
81
the nature of their usual occupations, internal mobility entails them
moving to occupations with little responsibility rather than greater
responsibility.
In this sector, the wage advantage for men is approximately
7.73%. The most of the wage gap was due to gender discrimina-
tion with values ranging from 77.78% to 83.6%. These results are in
line with previous studies in the hospitality sector. Campos-Soria
et al. (2009) estimated that gender discrimination ranged between
76.6% and 77.8%, assuming different returns and additional types
of gender segregation as control variables. The results suggest that
differences in returns to horizontal and vertical segregation, exter-
nal labour mobility, and marital status are the main sources of
gender discrimination. Vertical segregation reduces wage inequal-
ity because promotion increases women’s wages more than men’s
wages. However, horizontal segregation increases the wage differ-
ential because changing to a better paid functional area increases
menı́s wages more than women’s wages. Discrimination against
women that manifests as barriers to external mobility has a direct
effect on rewards. The data show not only that external mobility
is far greater than internal mobility in the hospitality industry, but
also that is the main cause of wage inequality between men and
women.
The results are subject to the usual limitations of a cross-
sectional database, which does not allow for controlling by
individual, time, or economic cycle effects. Sample selection bias
can generate inconsistent estimations of the wage equations
parameters. One study corrected for this type of bias by using the
method proposed by Heckman (1979), although the effect of using
this correction did not lead to statistical significance (Hernández,
1995). We have shown that most of the gender wage differential
among hospitality workers is due to discrimination, since there is a
disproportionate concentration of women in jobs with less respon-
sibility. However, this study does not address the issue of whether
such unequal distribution among men and women is due to gen-
der differences in preferences or family issues. Although the reason
women are concentrated in lower-paying jobs remains an enigma,
labour authorities should review the application of the Equal Pay
Act in this sector, placing special attention on barriers due to level
of education, hiring, and promotion. The data show the following:
women are not rewarded by specific training in tourism; there is a
disproportionate concentration of women in the worst-paid func-
tional areas; there are high levels of overeducation among women;
and women experience barriers to promotion to jobs with greater
responsibility. Thus, legislation should address these issues by facil-
itating access to the legal system and reducing the cost of litigation
as deterrents to discrimination.
References
Amuedo-Dorantes, C., De la Rica, S., 2006. The role of segregation and pay structure
on the gender wage gap: evidence from matched employer–employee data for
Spain. Contrib. Econ. Anal. Pol. 5 (1), 1–32.
Bayard, K., Hellerstein, J., Neumark, D., Troske, K., 2003. New evidence on sex
segregation and sex differences in wages from matched employee-employer
data. J. Labor Econ. 21 (4), 887–922.
Becker, G., 1964. Human Capital. National Bureau of Economic Research, New York.
Burns, P.M., 1993. Sustaining tourism employment. J. Sustain. Tourism 1 (2), 81–96.
Campos-Soria, J.A., Ortega-Aguaza, B., Ropero-García, M.A., 2009. Gender
segregation and wage difference in the hospitality industry. Tourism Econ. 15
(4), 847–866.
Campos-Soria, J.A., Marchante-Mera, A., Ropero-García, M.A., 2011. Patterns of
occupational segregation by gender in the hospitality industry. Int. J. Hosp.
Manage. 30, 91–102.
Delfim, L.D., Varejao, J., 2007. Employment, pay and discrimination in the tourism
industry. Tourism Econ. 13 (2), 225–240.
García-Pozo, A., Marchante-Mera, A., Sánchez-Ollero, J.L., 2011. Occupational
differences in the return on human capital in the Spanish travel agency and
hospitality industries. Tourism Econ. 17 (6), 1325–1345.
García-Pozo, A., Marchante-Mera, A., Sánchez-Ollero, J.L., 2014. Educational
mismatch and returns on human capital in the Spanish hospitality and travel
agency sectors. Tourism Econ. 20 (2), 337–354.
82 J.A. Campos-Soria et al. / International Journal of Hospitality Management 49 (2015) 73–82
García-Pozo, A., Campos-Soria, J.A., Sánchez-Ollero, J.L., Marchante-Lara, M., 2012.
The regional wage gap in the Spanish hospitality sector based on a gender
perspective. Int. J. Hosp. Manage. 31 (1), 266–275.
Gupta, N.D., Rothstein, D., 2005. The Impact of worker and establishment-level
characteristics on male-female wage differentials: evidence from Danish
matched employer–employee data. Rev. Labour Econ. Ind. Relat. 19 (1), 1–34.
Hakim, C., 1992. Explaining trends in occupational segregation: the measurement,
causes, and consequences of the sexual division of labour. Eur. Sociol. Rev. 8
(2), 127–152.
Heckman, J., 1979. Sample selection bias as a specification error. Econometrica 47,
153–161.
Hernández, P.J., 1995. Análisis empírico de la discriminación salarial de la mujer en
España. Investigaciones Económicas 19, 195–215.
IET, 2013. El turismo español en cifras 2012. Instituto de Estudios Turísticos.
Madrid.
INE, 2013. Cuenta Satélite del Turismo de España. Serie 2008-2011. Instituto
Nacional de Estadística. Madrid.
Jameson, S.M., 2000. Recruitment and training in small firms. J. Eur. Ind. Train. 24
(1), 43–49.
Jordan, F., 1997. An occupational hazard? Sex segregation in tourism employment.
Tourism Manage. 18 (8), 525–534.
Jovanovic, B., 1979. Job matching and the theory of turnover. J. Polit. Econ. 87 (5),
972–990.
Lee, C., Kang, S., 1998. Measuring earnings inequality and median earnings in the
tourism industry. Tourism Manage. 19 (4), 341–348.
Malo, M.A., Muñoz-Bullón, F., 2008. Working career progress in the tourism
industry: Temp-to-perm transitions in Spain. Business Economic Series 10,
Working Paper 08-35. Departamento de Economía de la Empresa. Madrid:
Universidad Carlos III.
Marchante, A.J., Ortega, B., Pagán, R., 2005. Educational mismatch and wages in the
hospitality sector. Tourism Econ. 11 (1), 103–117.
Marchante, A.J., Ortega, B., Pagán, R., 2007. An analysis of educational mismatch
and labor mobility in the hospitality industry. J. Hosp. Tourism Res. 31,
299–320.
McFadden, D., 1974. Conditional logit analysis of qualitative choice behaviour. In:
Zarembka, P. (Ed.), Frontiers of Econometrics. Academic Press, New York.
Mincer, J., 1974. Schooling, Experience And Earnings. Columbia University Press,
New York.
Neumark, D., 1988. Employerı́s discriminatory behaviour and the estimation of
wage discrimination. J. Hum. Resour. 23, 279–295.
Oaxaca, R.L., Ransom, M.R., 1994. On discrimination and the decomposition of
wage differentials. J. Econometrics 61, 5–21.
Pedraza, P., Muñoz de Bustillo, R., Villacampa, A., 2010. Determinantes de la
situación laboral en España: trabajar a tiempo parcial frente a otras situaciones
laborales. Cuadernos De Economía 33 (92), 71–104.
Riley, M., 1997. Human Resource Management in the Hospitality and Tourism
Industry, 2nd edition. Butterworth-Heinemann, Oxford.
Riley, M., Szivas, E., 2003. Pay determination. A socioeconomic framework. Annals
Tourism Res. 30 (2), 446–464.
Rosen, S., 1972. Learning and experience in the labor market. J. Hum. Resour. 7,
326–342.
Rosendaal, B.W., 2003. Dealing with part-time work. Pers. Rev. 32 (4), 474–491.
Salinas-Jiménez, M.M., Rahona-López, M., Murillo-Huertas, I.P., 2014. Gender wage
differentials and educational mismatch: An application to the Spanish case.
Appl. Econ. 45 (30), 4226–4235.
Sicherman, N., Galor, O., 1990. A theory of career mobility. J. Polit. Econ. 98,
160–192.
Sicherman, N., 1991. Overeducation in the labor market. J. Labor Econ. 9 (2),
101–122.
Simón, H.J., Ramos, R., Sanromá, E., 2008. Evolución de las diferencias salariales por
razón de sexo, 48. Revista de Economía Aplicada. XVI, pp. 37–68.
Thrane, C., 2008. Earnings differentiation in the tourism industry: gender, human
capital and socio-demographic effects. Tourism Manage. 29 (3), 514–524.
White, H., 1980. A heteroskedasticity-consistent covariance matriz estimator and a
direct test for heteroskedasticity. Econometrica 48, 817–830.
WTO, 2013. Panorama OMT del Turismo Internacional 2013. World Tourism
Organization, Madrid.