Original Scientific Article
Tourism and Economic Growth Nexus in Latin
America and Caribbean Countries:
Evidence from an Autoregressive Distributed
Lag Panel
José Alberto Fuinhas
ceber, Faculty of Economics, University of Coimbra, Portugal
fuinhas@uc.pt
Matheus Belucio
isag – European Business School and Research Group of isag (nidisag)
and cefage-ue, Department of Economics, University of Évora, Portugal,
matheus.silva@isag.pt
Daniela Castilho
Management and Economics Department, University of Beira Interior, Portugal
daniela.castilho@ubi.pt
JoanaMateus
Management and Economics Department, University of Beira Interior, Portugal
joana.mateus@ubi.pt
Rafaela Caetano
Management and Economics Department, University of Beira Interior, Portugal
rafaela.caetano@ubi.pt
This research focuses on tourism as a way to stimulate economic growth in Latin
America and the Caribbean countries. The impact of tourism on economic growth
was expected to have both short- and long-run effects. Panel autoregressive dis-
tributed lag, an econometric technique that allows for this temporal decomposition,
was used. The results for the twenty-two countries revealed that, in the short-run,
tourist capital investment per capita, tourist arrivals (number of persons), per capita
electricity consumption, and the real exchange rate were statistically significant and
had a positive impact on economic growth. In contrast, in the long-run, only tourist
arrivals and per capita electricity consumption proved to be positive drivers of per
capita economic growth. Policymakers should continue to develop and implement
measures to attract as many tourists as possible while promoting investment in the
tourism industry. However, they also need to pay attention to other economic sectors
so that their countries do not become extremely dependent on tourism activity.
Keywords: capital investment, tourism arrivals, economic growth, Latin America
and Caribbean, ardl
https://doi.org/10.26493/2335-4194.13.21-34
Academica Turistica, Year 13, No. 1, June 2020 | 21
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Introduction
Although the economic activity of Latin America and
Caribbean countries is still recovering from the im-
pacts of various economic and social crises, the Inter-
national Monetary Fund (imf) stated that after 2017
the region’s economic growth increased by 1.3, 1.6
in 2018, and 2019 it is expected to increase by 2.6 (In-
ternationalMonetary Fund, 2018). These results could
mean that these countries can rapidly increase their
growth rates. The primary motivation for the achieve-
ment of this study was the fact that tourism is a sector
in development and has an essential role in the eco-
nomic growth of this region. Thus, corroborating with
World Travel Market (2018), the number of foreign
tourists arriving in Latin America increased by 6 in
the previous five years. Moreover, the World Travel
and TourismCouncil (2018) estimates point to the fact
that the travel and tourism sector recently contributed
to 15.2 of the Caribbean Gross Domestic Product, re-
gion which is included in the group of countries that
we will use in our investigation.
Given the facts previously stated, the characteris-
tics of this region and the particular interest points,
tourism becomes quite relevant. Being a fast-growing
sector, it is crucial to verify if more tourists and in-
vestment (as well as, their efficiency) lead to an in-
crease in economic benefits of Latin American and
Caribbean countries. Thus, this theme should con-
tinue to be studied.
The impact of tourism on economic growth is an
extensively explored theme in the economic growth
literature. However, few studies use panel methodol-
ogy to study the impacts of this sector on the growth
of the Latin America and Caribbean region.
In our study, we used 22 countries from the Latin
America andCaribbean region,with annual data rang-
ing from 1995 to 2014. The autoregressive distributed
lag (ardl)model was used in our empirical investiga-
tion mainly because it supports variables with differ-
ent orders of integration and gives robust results with
small samples. Though the ardl model, we evaluate
the impacts that tourism intensity and tourism capi-
tal investment have on the economic growth of Latin
America and Caribbean countries on both the short
and that long run. To reach this objective, we used
annual data on Gross Domestic Product per capita
(gdppc), which is our proxy for economic growth,
and on tourism arrivals per capita (tapc) and tourism
capital investment per capita (tipc) in order to repre-
sent the tourism sector.
The main goal of this study is to answer the cen-
tral question: ‘What are the impacts of tourism inten-
sity and tourism capital investment on the economic
growth of Latin America and Caribbean countries?’
Given the central question of our study, we can con-
struct the two following hypotheses.
h1 Tourism intensity has a positive impact on eco-
nomic growth, given that it contributes to em-
ployment creation and stimulates the economic
activity of the Latin America and Caribbean
countries.
h2 Capital investment has a positive impact on
economic growth, given that it contributes to
the construction of new infrastructure, techno-
logical progress, and innovation in Latin Amer-
ica and Caribbean countries.
In this research, wewill attempt to confirm (or not)
the validity of these hypotheses as we simultaneously
attempt to contribute to the enlargement of the litera-
ture on this field.
This study is organised as follows. The second sec-
tion presents literature reviews about the tourism-
economic growth nexus. The third describes the data,
methodology, and preliminary tests. The fourth sec-
tion presents the results and discussion, and the fifth
section concludes the study.
Literature Review
In this section, essential aspects will be discussed in
the literature on tourism economics, focusing on the
way tourism relates to the economy and addressing
specific aspects of tourism in the Latin American and
Caribbean region.
Revisiting Tourism and Economy
The relationship between tourism development and
economic growth has been studied widely in recent
years (e.g., Cannonier & Burke, 2019; Belucio et al.,
2018; Brida, Lanzilotta & Pizzolon, 2016; Du et al.,
22 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
2016; Cárdenas-García et al., 2015; Tugcu, 2014) given
that the results from this relationship can help some
countries to develop effective growth strategies for
their economies.
The supply of foreign currency, the promotion of
investment, in both infrastructure and human capi-
tal, and the jobs that this sector creates are some of
the significant benefits from international tourism
that can produce positive effects on a country’s eco-
nomic growth.Moreover, following Blake et al. (2016),
tourism has an essential role on the increase of the
average income of a country, as well as in the in-
crease of both the efficiency and competitiveness of
the economies.
Malta et al. (2019) analysed the context that enabled
the creation of a vision that attributes to tourism the
capacity to reduce poverty. According to the World
Tourism Organization (2015), in 2014, one in every
eleven jobs around the world was created by the tour-
ism sector, which demonstrates the weight that this
sector has on the worldwide economy.
Given the facts previously stated, it is natural that
the relationship between tourism and growth has an
extensive branch of literature about it. Lanza and Pigli-
aru (2000) were the pioneers of the investigation of
the relationship between these two variables. In their
work, they concluded that the countries specialised in
tourism shared features, such as tourist destinations of
small geographical size, where the average per capita
income proliferated.
Moving forward, the analysis of the relationship
between tourism and economic growth has at least
four hypotheses that can be easily identified in the lit-
erature: growth hypothesis also called the tourism-led
growth hypothesis (tlgh or tlg): conservation hy-
pothesis; feedback hypothesis; and neutrality hypoth-
esis (e.g., Dogru & Bulut, 2018). These hypotheses ap-
pear in the majority of the causality tests related to
economic growth, mainly in energy economics. Nev-
ertheless, this group of hypotheses can be reformu-
lated and used in tourism analysis.
The tlgh, as the name implies, state that tourism
development stimulates economic growth: the tourist
arrivals and the revenues generated by the tourism sec-
tor have a positive impact on economic growth. This
hypothesis is supported by the majority of the authors
that focus their works on the assessment of this rela-
tionship (e.g., Shahzad et al., 2017; Tugcu, 2014;Husein
& Kara, 2011; Cortes-Jimenez & Pulina, 2010).
Regarding the conservation hypothesis (e.g., Aslan,
2014; Payne &Mervar, 2010), while this hypothesis as-
serts that the economic output of a country can induce
tourism development, it also suggests that deteriora-
tion on the economic performance of a country can
significantly reduce its tourism demand.
Concerning the feedback hypothesis (e.g., Rivera,
2017; Al-mulali et al., 2014;Massidda&Mattana, 2013),
it considers economic growth and tourism develop-
ment to be complementary and strongly dependent.
This hypothesis is the same as saying that economic
growth promotes tourism development as well as the
other way around.
Finally, the neutrality hypothesis (e.g., Katircioglu,
2009) suggests that there is no relationship between
tourism development and economic growth: they are
entirely independent. This hypothesis indicates that,
for example, strategies for tourism development (e.g.,
investing in the tourism sector) do not produce direct
effects on economic growth.
Besides these four hypotheses, there is an addi-
tional one: the curse hypothesis, or the beach disease
effect (e.g., Holzner, 2011). This hypothesis can be de-
fined as follows: countries in which the tourism sector
plays a significant role in their economies (tourism-
dependent countries) tend to grow less than the others
do.
Turning to the methodological part of the works
on the tourism-growth nexus, in the literature, many
variables were used as tourismproxies. Themost com-
monly used are international tourism revenues (e.g.,
Durbarry, 2004; Balaguer & Cantavella-Jordà, 2002),
number of tourist arrivals (e.g., Zortuk, 2009; Gun-
duz & Hatemi-J, 2005), tourism specialization (e.g.,
Algieri, 2006), tourism industries (e.g., Tang& Shawn,
2009), and tourism spending (e.g., Nissan et al., 2011),
for example. As expected, the variable Gross Domestic
Product is the one that researchers use the most often
to measure economic growth.
Regarding the empirical methodologies that are
used to investigate the relationship between tourism
Academica Turistica, Year 13, No. 1, June 2020 | 23
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
and economic growth, it can be emphasized that they
differ from author to author. However, there are two
main methodological approaches: panel data estima-
tions, and time-series estimations. Brida et al. (2016)
made a detailed review of the empirical methods that
were applied in the literature close to this theme, and
their advantages and disadvantages.
The panel data models are frequently preferred
because they allow doing a simultaneous analysis of
the cross-sectional and temporal dimensions. The
panel Granger causality techniques (e.g., Belucio et
al., 2018; Al-mulali et al., 2014) and the autoregressive
distributed lag model (ardl) (e.g., Katircioglu, 2009)
are some of the estimation methods that are more fre-
quently used in this type of studies.
The Region
LatinAmerica has been experiencing significant chan-
ges in recent decades (Bianchi et al., 2018). Tourism
has grown in most Latin American and Caribbean
countries. Researchers and policymakers have long
recognised the significance of tourism to theCaribbean
region (Cannonier & Burke, 2019)
The region has two of the seven natural wonders
of the world, and three of the seven wonders of the
modern world. Tourism in the region has diversified
effects, whethermicro or macroeconomics. Garza and
Ovalle (2019) argue that tourism-driven development
affects the spatial distribution of prices and increasing
daily transportation difficulties.
Focusing on the Latin America and Caribbean
countries (lac), we can refer that for these countries
the tourism literature is quite extensive (e.g., Belucio
et al., 2018; Risso & Brida, 2008; Brida et al., 2008;
Eugenio-Martin et al., 2004). The conclusions of the
works focused on the relationship between tourism
and growth in this region, or in some countries of
the lac, predominantly support the tlgh. Before we
conclude, we also should refer that the reasons cited
in the literature to the differences in the study’s results
are mainly the fact that authors usually apply different
empirical methodologies, and chose different periods
and samples to be analysed (e.g., Dogru&Bulut, 2018).
Even thoughmost of the studies show that tourism
development has positive impacts on the economic
output of the countries, the results are far from conclu-
sive, and for that reason, we support the idea that this
relationship should continue to be extensively studied.
The assessment of the impacts that tourism have
on growth is especially crucial for the case of Latin
America and Caribbean countries because they have a
set of characteristics (e.g., cultural and natural wealth)
that make them a choice destination for tourists from
all over the world. However, tourism safety in Latin
America has not evolved to the same level in all Latin
American countries (Maximiliano, 2014).
Some of the critical factors that can deter tourists
from a destination are the security of the destination
and the exchange rate. Regarding the safety of tourists
in Latin America, the central issue is related to lo-
cal crime (Maximiliano, 2014). However, the devel-
opment of sound public regulation can generate eco-
nomic growth and benefits for tourism agents, which
is reflected in improvements for the population (Belu-
cio et al., 2018) and tourists.
The exchange rate plays an essential role in the lives
of underdeveloped or in the development of tourist
destinations. The inflow of foreign capital is responsi-
ble for economic growth, but policymakers often ne-
glect the exchange rate policies (Dogru et al., 2019),
which can have a significant impact on the trade bal-
ance. It is also known that the real effective exchange
rate has significant effects on economic growth (Lee &
Chang, 2008) and that exchange variation can benefit
or hurt a tourist destination.
Data andMethodology
Our study is focused on the assessment of the impacts
of tourism on the economic growth of a group of Latin
America and Caribbean: Argentina, Bolivia, Brazil,
Chile, Colombia, Costa Rica, Cuba, Dominican Re-
public, Ecuador, El Salvador, Guatemala, Haiti, Hon-
duras, Jamaica,Mexico, Nicaragua, Panama, Paraguay,
Peru, Trinidad and Tobago, Uruguay and Venezuela.
For the present investigation, we will use annual data
from 1995 to 2014. Both the time horizon and countries
were chosen, given the available data. In this study, we
used stata 15.0 to perform our econometric analy-
sis. In Table 1 the name, definition, and source of our
variables are presented.
24 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 1 Variables Description
Variable Definition Source
gdp Gross Domestic Product in
a constant local currency
unit
World Bank
p The total population in the
total number of persons
World Bank
epc Electric power consumption
in gwh
World Bank
ta Tourism arrivals in the
number of persons
World Bank
gdp_us Gross Domestic Product in
constant 2010 us$
World Bank
ti Capital Investment in a
constant local currency unit
World Travel &
Tourism Council
tx Real exchange rate Author’s calcu-
lation from the
World Bank
The dependent variable will be Gross Domestic
Product in constant local currency (gdp), our proxy
for economic growth.
To measure tourism intensity, the ratio of the tour-
ism arrivals (ta), by the total population (p) was used.
The tourism capital investment (ti), which represents
the capital investment spending by all industries di-
rectly involved in travel and tourism, is another of our
interest variables and will also be divided by the to-
tal population (p). We choose the electric power con-
sumption (epc) as our control variable because the
energy use of a country is highly correlated with its
economic growth (e.g., Santiago et al., 2020); further-
more, energy can contribute to the three dimensions
of development: social, economic and human (e.g.,
Malaquias et al., 2019). The Gross Domestic Prod-
uct was also retrieved in constant 2010 us$, in or-
der to calculate the real exchange rate (tx) through
the ratio of Gross Domestic Product in constant local
currency by the Gross Domestic Product in constant
2010 us$.
The transformation of the variables into per capita
values is essential because it eliminates the dimen-
sional distortions caused in the model’s estimations by
the variables in levels. gdp, epc, ta, gdp_us, and p,
were all retrieved from the World Bank, while ti was
retrieved from the World Travel & Tourism Council.
We will use the autoregressive distributed lag (ar-
dl) model in the form of an Unrestricted Error Cor-
rection Mechanism (uecm). This methodology gives
the dynamic effects of the variables, allowing us to
make a distinction between the Granger causality in
short and the long-run. Moreover, it is robust to the
presence of endogeneity, and when a determined co-
efficient is statistically significant, it is equivalent to
the Granger causality testing (Menegaki et al., 2017;
Jouini, 2015). Additionally, it deals with cointegration
and supports the inclusion of variables with different
orders of integration (I(0), I(1), and fractionally inte-
grated variables) in the same estimation. The variables
were transformed into natural logarithms (‘L’) and first
differences (‘D’). The ardl model specification is the
following:
lgdppcit = α1i + δ1itrend + β1i1lgdppcit−1
+ β1i2ltapcit + β1i3ltapcit−1
+ β1i4ltipcit + β1i5ltipcit−1
+ β1i6lepcit + β1i7lepcit−1
+ β1i8txit + β1i9txit−1 + β1it. (1)
To explain the dynamic relationships between our
variables, we reparametrized equation (1) into the fol-
lowing specification:
dlgdppcit = αi + β2i1dltapcit + β2i2dltipcit
+ β2i3dlepcit + β2i4dltxit
+ γ2i1lgdppcit−1 + γ2i2ltapcit−1
+ γ2i3ltipcit−1 + γ2i4lepcit−1
+ γ2i5txit−1 + ε2it . (2)
A series of diagnostic tests before the estimation are
necessaries to validate that the choice of method was
accurate. In addition, other tests and statistics need to
be verified after model estimation to make sure that it
meets mandatory econometric requirements in panel
analysis (e.g., Dogru et al., 2019; Santiago et al., 2020;
Marques et al., 2017; Fuinhas & Marques, 2012; Katir-
cioglu, 2009). Every test and statistics of the method
used will be presented.
In sequence, the characteristics of the series through
Academica Turistica, Year 13, No. 1, June 2020 | 25
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 2 Descriptive Statistics and Cross-Sectional Dependence
Variables Descriptive statistics Cross section dependence (cd)
Obs Mean Std. dev. Min. Max. cd-test Corr Abs(corr)
lgdppc  . . . . .*** . .
ltapc  –. . –. –. .*** . .
ltipc  –. . –. –. .*** . .
lepc  . . . . .*** . .
ltx  . . –. . –. . .
dlgdppc  . . –. . .*** . .
dltapc  . . –. . .*** . .
dltipc  . . –. . .*** . .
dlepc  . . –. . .*** . .
dltx  . . –. . –. –. .
Notes To achieve the results of descriptive statistics and to test the presence of cross-section dependence, the Stata com-
mands sum and xtcd, respectively, were used. The cd test hasN(0, 1) distribution under the h0: cross-section independence;
*** denote statistical significance at 1 level.
the descriptive statistics as well as the results from the
cross-section dependence test are presented.As can be
observed, gdppc has one less observation (data on
gdp fails for Haiti in 1995), but that is not a problem,
because stata 15.0 can correct this issue and contin-
ues to assume the panel to be a strongly balanced one.
It is also possible to observe one less observation on
tx because the variable was calculated using the gdp,
and it is not a concern due to the same explanation. As
previously stated, in Table 2, the results from the cross-
section dependence test can be observed, where it can
be concluded that cross-section dependence is present
in all variables, except in real exchange rate (tx).
Next, the correlation matrices and variance infla-
tion factor (vif) statistics are examined. The correla-
tion matrix was used to check the degree of correla-
tion that exists between the variables, while the vif
statistics was used to test for the presence of multi-
collinearity. The results of the correlation matrix only
indicated the existence of a high level of correlation
between ltipc and lgdppc, which is not a concern,
given that the high correlation is with the dependent
variable. A similar situation (high correlation between
the ltx and lgdppc) is detected; again, this does not
cause a problem for the estimation due to the same rea-
son. The lower vif and mean vif values prove that
Table 3 Correlation Matrices and vif Statistics
lgdppc ltapc ltipc lepc ltx
lgdppc .
ltapc . .
ltipc . . .
lepc . . . .
ltx . . . . .
vif* n.a. . . . .
dlgdppc dltapc dltipc dlepc dltx
dlgdppc .
dltapc . .
dltipc . . .
dlepc . . . .
dltx . –. . . .
vif** n.a. . . . .
Notes *Mean vif 2.12. **Mean vif 1.04.
multicollinearity is not a problem for this paper’s esti-
mation. Details are in the Table 3.
Because cross-sectional dependence seems not to
be present on the real exchange rate (tx), the 1st gener-
ation panel unit root tests will also be computed. Next,
the results of theMaddala andWu test are presented in
26 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 4 Maddala and Wu Panel Unit Roots Test (mw)
Variable wb (Zt-bar)
Without trend With trend
lgdppc .*** .***
ltapc .* .
ltipc .* .***
lepc . .
ltx .*** .***
dlgdppc .*** .***
dltapc .*** .***
dltipc .*** .***
dlepc .*** .***
dltx .*** .***
Notes *, *** denote statistical significance at 10 and 1
level, respectively; Maddala and Wu (1999) Panel Unit Root
Test (mw) assumes that cross-sectional independence, and
h0: series is i(1); to compute this test, the Stata command
multipurt was used.
Table 4. As the test of Maddala and Wu (1999) shows
that there are variables presenting cross-sectional de-
pendence, the veracity of the results is compromised
for them. Thus, only the order of integration of the tx
variable will be analysed. The result seems to indicate
that the variable is I(0).
To see the order of integration of the remaining
variables, the 2nd generation unit root tests, namely
the augmented cross-sectional ips (cips) test by (Pe-
saran, 2007) were computed. This test was used be-
cause the presence of cross-sectional dependence was
registered in most of the variables, and the 1st gen-
eration panel unit root tests turned out to be ineffi-
cient in these cases. The results of the cips test show
that some variables are I(1) and others are I(0), which
is not a problem because the ardl model supports
these two levels of integration. These results confirm
that the ardl methodology is the best approach for
the study (Table 5).
The Hausman test confronts random and fixed ef-
fects, andwhen the structure of the data is in the panel,
it is necessary to test for the individual effects. In se-
quence, the results of the Hausman test are presented
and, as we can be observed, the test rejects the null hy-
Table 5 Panel Unit Root test (cips)
Variable cips (Zt-bar)
Without trend With trend
lgdppc –. .
ltapc –.*** –.
ltipc –.*** –.***
lepc –. .
ltx –.*** –.***
dlgdppc –.*** –.***
dltapc –.*** –.***
dltipc –.*** –.***
dlepc –.*** –.***
dltx –.*** –.***
Notes *** denote statistical significance at 1 level; Pesaran
(2007) Panel Unit Root Test (cips) assumes that cross-
sectional dependence is in the form of a single unobserved
common factor and h0: series is i(1); to compute this test,
the Stata commandmultipurt was used.
Table 6 Hausman Test
Test fe vs re
Hausman test χ2() = .***
Notes *** denotes significance at the 1 level; in bothmod-
els, theHausman test was performedwith the sigmamore op-
tion. h0: random effects are the most appropriate.
pothesis. This result led us to conclude that the fixed
effects model is the proper specification for our es-
timation: the countries’ individual effects are signif-
icant. In this estimation, the sigmamore option was
used, which is a recurrent option in previous studies
(e.g., Özokcu & Özdemir, 2017) (Table 6).
After theHausman test, with the results pointing to
the use of the fixed effects model, the next step is the
execution of a group of specification tests. The results
of the pre-tests still reveal details of the nature of the
variables, information useful for models’ estimation.
Results and Discussion
To test for the presence of heteroscedasticity, we com-
puted the modified Wald Test (null hypothesis: Ho-
moscedasticity). The Pesaran test (null hypothesis:
Academica Turistica, Year 13, No. 1, June 2020 | 27
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 7 Specification Tests
Test Statistics
Modified Wald test .***
Pesaran’s test .***
Wooldridge test .***
Notes h0 of Modified Wald test: σ(i)2= σ2 for all i; H0 of
Pesaran’s test: residual are not correlated; h0 ofWooldridge
test: no first-order autocorrelation; *** denotes statistical sig-
nificance at 1 level.
residuals are not correlated and follow a normal distri-
bution) to check for the presence of contemporaneous
correlation was used. The Breush-Pagan Lagrangian
multiplier test was also used to test if the variances
across individuals are not correlated. In the present
case, this test could not be applied because the number
of countries in that sample is larger than the number
of years in the study. Lastly, the Wooldridge test was
used for autocorrelation to assess for the presence of
serial correlation in our model.
The results from the previously mentioned tests
are presented in Table 7, showing that heteroscedas-
ticity, contemporaneous correlation, and first-order
autocorrelation are all present in the model. All the
statistics reject the null hypothesis of the respective
specification tests.
Given these results, the Driscoll and Kraay (1998)
estimator is the most appropriate estimator to use in
estimations, because the standard errors produced by
the estimator are robust to disturbances being cross-
sectional dependent, heteroskedastic and autocorre-
lated up to some lag.
In this model, both the trend, the tourism capi-
tal investment per capita, and the real exchange rate
(both ti and tx on the long-run) were statistically
insignificant and were thus retrieved from the model.
After these conclusions, equation (2) was replaced by
equation (3), which represents thismore parsimonious
model.
dlgdppcit = αi + β3i1dltapcit + β3i2dltipcit
+ β3i3dlepcit + β3i4dltxit
+ γ3i1lgdppcit−1 + γ3i2ltapcit−1
+ γ3i3lepcit−1 + εit. (3)
Table 8 Estimation Results
(Dependent Variable: dlgdppc)
Variable fe fe-dk
Constant .*** .***
dltapc .*** .***
dltipc .*** .***
dlepc .*** .***
dltx .*** .***
lgdppc (–) –.*** –.***
ltapc (–) .*** .***
lepc (–) .*** .*
Diagnostic statistics
N  
R2 . .
F F(, ) = .*** F(, ) = .***
Notes ***, * denote statistical significance at 1 and 10
level, respectively; to estimate the models, the Stata com-
mand xtscc was used.
Specification tests were remade for the parsimo-
nious model, and the results were in line with the
previous ones (presence of heteroscedasticity, auto-
correlation. and contemporaneous correlation in the
model). The results of the estimations are presented in
detail in Table 8. The results show that, in the short-
run, the tourism intensity, the tourism capital invest-
ment per capita, the electric energy consumption per
capita, and the real exchange rate are all positive and
statistically significant. Table 8 also shows a positive
and statistically significant impact of both tourism in-
tensity and electric power consumption on economic
growth in the long-run. As we previously stated, the
tourism capital investment and real exchange rate
failed to show a statistically significant impact in the
long-run and, for that reason, it was excluded from the
estimation.
The long-run elasticities are not displayed in Ta-
ble 8 because they had to be calculated through the ra-
tio between the variable’s coefficient and the lgdppc
coefficient, both lagged once, and this ratio was mul-
tiplied by –1. In Table 9, the impacts (short-run), elas-
ticities (long-run), and the adjustment speed of the
model (ecm) are shown.
28 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 9 Elasticities and Speed of Adjustment
(Dependent Variable: dlgdppc)
Variable fe fe-dk
Short-run impacts
dltapc .*** .***
dltipc .*** .***
dlepc .*** .***
dltx .*** .***
Long-run (computed) elasticities
ltapc .*** .***
lepc .*** .***
Speed of adjustment
ecm –.*** –.***
Notes *** denote statistical significance at 1 level, the
ecm denotes the coefficient of the variable lgdppc lagged
once.
From Table 9, it can be seen that the Latin Amer-
ica and Caribbean countries’ economic growth was
positively affected by the tourism arrivals per capita
(tourism intensity) and by the electric power con-
sumption per capita, both in the short and long runs,
while the positive effects of tourism capital investment
per capita and real exchange rate were only detected
in the short run.
LatinAmerica andCaribbean countries suffer from
serious political, economic and social problems and
therefore, once these problems had an impact on the
economic growth of these countries, we considered
the relevant shocks, which affected their economies
between 1995 and 2014.
In 1997, the Mexican government adopted the Na-
tional Program for Development Finance (npfdf). In
Uruguay, in 2002, a bank crisis occurred, due to the
country’s over-dependence on Argentina, which was
also in depression. This depression was mainly due to
currency devaluation. In the Dominican Republic, in
2003, a financial crisis was caused by bank failure.
In Venezuela, the oil strike in 2002–2003 followed
in 2004 with an impressive rise in the oil prices. Trini-
dad and Tobago are very dependent on exports; in
2003m this country registered a massive increase in
gdp, which could be associated with the Venezuelan
Table 10 Estimation Results (Corrected for Shocks,
Dependent Variable: dlgdppc)
Variable fe fe-dk
Constant .*** .***
dltapc .*** .***
dltipc .*** .***
dlepc .*** .***
dltx .*** .***
lgdppc (–) –.*** –.***
ltapc (–) .*** .***
ltipc (–) . .**
lepc (–) .*** .**
arg –.*** –.***
arg –.*** –.***
cb .*** .***
cb .*** .***
rd –.*** –.***
h –.** –.***
mex .** .***
mex –.*** –.***
tt .*** .***
tt .*** .***
tt –.*** –.***
ur –.*** –.***
ven –.*** –.***
ven –.*** –.***
ven .*** .***
Diagnostic statistics
N  
R2 . .
F F(, ) = .*** F(, ) = .***
Notes *** and ** denote statistical significance at 1 or 5
level, respectively; to estimate the models, the Stata com-
mand xtscc was used.
instability in the same year, which led to a search for a
new hydrocarbon exporting country, which benefitted
Trinidad and Tobago.
In 2006, Trinidad and Tobago, due to a rise in the
oil and gas prices and an increase in the foreign di-
rect investment (fdi), expanded their energy sector.
Academica Turistica, Year 13, No. 1, June 2020 | 29
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Table 11 Impacts, Elasticities and Speed of Adjustment
(Model Corrected for Shocks, Dependent
Variable: dlgdppc)
Variable fe fe-dk
Short-run impacts
dltapc .*** .***
dltipc .*** .***
dlepc .*** .***
dltx .*** .***
Long-run (computed) elasticities
ltapc .*** .***
ltipc . .**
lepc .*** .***
Speed of adjustment
ecm –.*** –.***
Notes *** denote statistical significance at 1 level, the
ecm denotes the coefficient of the variable lgdppc lagged
once.
In 2005, Cuba had a development of the tourism sec-
tor and was registered a reduction in the unemploy-
ment rate. In 2006, the highest economic growth in
the history of Cuba happened as a result of Cuba’s
so-called energy revolution (e.g., Suárez et al., 2012).
Other shocks that were considered were in Argentina,
Haiti, Mexico and Trinidad and Tobago (all in 2009)
and that can be due to the financial crisis of 2008 fol-
lowed by a global recession. Details in Table 10.
What was said previously about some economic
problems in these countries indicates the existence of
outliers in Argentina (2002), Cuba (2005, 2006), the
Dominican Republic (2003), Mexico (1997), Trinidad
and Tobago (2003, 2006), and Venezuela (2002, 2003,
2004). To control the detected outliers, dummies were
added on the model to represent these events and cor-
rect them. Dummies arg2002, arg2009, rd2003,
h2009, mex2009, tt2009, ur2002, ven2002, and
ven2003 represent a break, while cb2005, cb2006,
mex1997, tt2003, tt2006 and ven2004 represents
a peak. In Table 11, the impacts, elasticities and speed
of adjustment of the model are shown.
From Table 11, it can be seen that the Latin Amer-
ica and Caribbean countries economic growth was
positively affected by the tourism arrivals per capita
(tourism intensity) and by the electric power con-
sumption per capita, both in the short and long run.
After the correction of the shocks, the tourism invest-
ment per capita has become statistically significant,
not only on the short-run but also in the long-run. In
addition, it had a positive impact on economic growth,
becoming one of its main drivers. The real exchange
rate has a significant and positive impact on economic
growth but only in the short-run.
Regarding the ecm, from Table 11, it can be seen
that its coefficient is negative and statistically signif-
icant, which indicates the presence of long-memory
between the variables. This value represents the speed
of adjustment of the model, i.e., the speed at which
the dependent variable returns to equilibrium after
changes in our independent variables. As can be ob-
served, the speed of adjustment of the model is rela-
tively slow.
The positive impacts of the electric power con-
sumption per capita on the economic growth of these
countries, both in the short and long-run, were ex-
pected, given that energy is seen as a driving force
for growth (e.g., Hatemi-J & Irandoust, 2005). Addi-
tionally, it has high explanatory power in empirical
growth models. Moreover, many authors consider en-
ergy variables crucial to explain countries’ economic
growth (e.g., Toman & Jemelkova, 2003). The real ex-
change rate, as previously stated, had a positive im-
pact on the economic growth in the short run. The
importance of the exchange rate to the policy and eco-
nomic growth could benefit the countries that were in
the early stages of economic development (Habib et
al., 2017). Thus, because the countries used in this in-
vestigation are developing countries, this impact was
expected. In the long run, with countries becoming
more developed and prosperous, the real exchange
rate could become irrelevant to growth (Aghion et al.,
2009).
Given these results, the policymakers from Latin
America and the Caribbean should be cautious in
the adoption of energy conservation policies, since
the economic output of these countries seems to be
strongly linked with energy consumption, in the pres-
ent case, with the electric power consumption per
30 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
capita. Measures that lead to a reduction in its con-
sumption appear to be able to affect the economic
growth of Latin America and Caribbean countries ad-
versely.
Regarding the central question of the present study,
it can be seen that both variables (tourism intensity
and tourism capital investment) seem to have had a
positive impact on growth, which confirms both of
the hypotheses. The results of this study also corrob-
orate those of other authors that studied the relation-
ship between the tourism sector and economic growth
for some countries from this region (e.g., Shahzad et
al., 2017; Tang & Abosedra, 2014; Amaghionyeodiwe,
2012).
Given these results, we think that the countries
from our sample should continue to attract as many
tourists as possible at the same time, while the indus-
tries directly involved in travel and tourism should
continue to increase the levels of their investments,
given that both factors have a positive impact on eco-
nomic growth. This is congruent with the findings of
Du et al. (2016) that tourism’s contribution to the long-
run growth of an economy comes through its role as
an integral part of a broader development strategy.
Conclusion
In order to answer to the central question of this
study, the autoregressive distributed lag (ardl)model
was used to assess the impacts, in both the short and
long-run, of tourism on the economic growth of 22
Latin America and Caribbean countries. The specifi-
cation tests showed that cross-sectional dependence,
heteroscedasticity, contemporaneous correlation, and
first-order autocorrelation were present in the model,
which led to the Discroll Kraay estimator with fixed
effects being used. The Error Correction Mechanism
(ecm) is statistically significant and negative, which
indicates the presence of cointegration/long-memory
relationships between the variables in the study.
From the results, it is possible to observe that, in
the short run, tourism intensity, capital investment,
electric power consumption, and real exchange rate
have a positive and significant impact on the economic
growth of the Latin America and Caribbean countries,
with the electric power consumption per capita being
themain driver of the growth. In the long-run, all vari-
ables were shown to be significant and have a positive
impact on growth. The tourism arrivals and electric
power consumption have been revealed to be the prin-
cipal drivers of economic growth in this region.
Therefore, the tourism intensity and capital invest-
ment, both on short and long-run, had a positive im-
pact on the economic growth of the Latin America
and Caribbean countries, which supports Hypotheses
1 and 2.
The main finding of this investigation is that once
that tourism has a positive impact on the economic
growth of this region, which means that an increase
on the tourism intensity leads an increase on the eco-
nomic growth, this region should increase the level
of investment in this sector. The policymakers of the
Latin America and Caribbean region should continue
to develop measures aimed to attract as many tourists
as possible while simultaneously promoting the in-
vestment in their travel and tourism industries. The
country’s economies have to invest more in human
capital directly involved with the tourism sector and
invest more in marketing to promote the region of
Latin America and the Caribbean in addition to other
economic sectors.
When a tourist chooses one destination, the ma-
jority of them (or all of them) do so considering the
economic situation, the level of security, and the pub-
lic health conditions of the region. Consequently, the
policymakers should increase the level of the invest-
ments in healthcare (both to residents and tourists),
which could happen through international partner-
ships with tourism agencies, for example, and should
also increase the security in the region.
However, they also must pay attention to the other
economic sectors so that their countries do not be-
come extremely dependent on tourism activity. Exces-
sive investment in the tourism sector, while neglecting
the other sectors of the economymay lead these coun-
tries to a ‘deindustrialisation’ situation.
The use of energy consumption or electric power
consumption directly related to tourism should be in-
cluded in further research because it is a limitation of
this investigation, as is the temporal horizon that ends
in 2014. Another limitation of the study is analysing
Academica Turistica, Year 13, No. 1, June 2020 | 31
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
the exchange rate behaviour with on linear analysis,
which is different from what is commonly addressed
in the literature, non-linearmethods (e.g., Dogru et al.,
2019; Irandoust, 2019).
We note that another gap in the tourism litera-
ture and economic growth that may be incorporated
in future research: the inclusion of exogenous vari-
ables representing instability (e.g., political instabil-
ity). Thus, allowing a more robust empirical approach
to the current problems of the countries of the re-
gion (e.g., Venezuela, Brazil, Argentina, Bolivia) could
guarantee greater veracity of the results (e.g., Arslan-
turk et al., 2011; Chen & Chiou-Wei, 2009).
Acknowledgments
ceber, r&d unit funded by national funds through
fct –Fundação para aCiência e a Tecnologia, ip, project
uidb/05037/2020; nece, r&d unit funded by national
funds through fct – Fundação para a Ciência e a Tec-
nologia, ip, project uidb/04630/2020; cefage, r&d
unit funded by national funds through fct – Fundaçâo
para a Ciência e a Tecnologia, ip, project uidb/04007/
2020.
References
Aghion, P., Bacchetta, P., Rancière, R., & Rogoff, K. (2009).
Exchange rate volatility and productivity growth: The
role of financial development. Journal of Monetary Eco-
nomics, 56(4), 494–513.
Algieri, B. (2006). International tourism specialisation of
small countries. International Journal of Tourism Re-
search, 8(1), 1–12.
Al-mulali, U., Fereidouni, H. G., Lee, J. Y.M., &Mohammed,
A. H. (2014). Estimating the tourism-led growth hypoth-
esis: A case study of the Middle East countries.Anatolia,
25(2), 290–298.
Amaghionyeodiwe, L. A. (2012). Research note: A causality
analysis of tourismas a long-run economic growth factor
in Jamaica. Tourism Economics, 18(5), 1125–1133.
Arslanturk, Y., Balcilar, M., & Abidin, Z. (2011). Time-vary-
ing linkages between tourism receipts and economic
growth in a small open economy. Economic Modelling,
28(1–2), 664–671.
Aslan, A. (2014). Tourism development and economic grow-
th in the Mediterranean countries: Evidence from panel
Granger causality tests. Current Issues in Tourism, 17(4),
363–372.
Balaguer, J., & Cantavella-Jordá, M. (2002). Tourism as a
long-run economic growth factor: The Spanish case.Ap-
plied Economics, 34(7), 877–884.
Belucio, M., Antunes, J., Fernandes, F., Fuinhas, J. A. &Mar-
tin,M. B. L. (2018, November). Revisiting the tourism on
Latin America: A panel analysis. Paper presented at Se-
meAd 2018 – xxi Seminário emAdministração da Uni-
versidade de São Paulo.
Bianchi, C., Mingo, S., & Fernandez, V. (2018). Strategic
management in LatinAmerica: Challenges in a changing
world. Journal of Business Research, 105, 306–309.
Blake, A., Sinclair, M. T., & Soria, J. A. C. (2006). Tourism
productivity: Evidence from the United Kingdom. An-
nals of Tourism Research, 33(4), 1099–1120.
Brida, J. G., Cortes-Jimenez, I., & Pulina, M. (2016). Has the
tourism-led growth hypothesis been validated? A litera-
ture review. Current Issues in Tourism, 19(5), 394–430.
Brida, J. G., Lanzilotta, B., & Pizzolon, F. (2016). Dynamic
relationship between tourism and economic growth
in mercosur countries: A nonlinear approach based
on asymmetric time series models. Economics Bulletin,
36(2), 879–894.
Brida, J. G., Lanzilotta, B., & Risso, W. A. (2008). Turismo y
crecimiento económico: El caso de Uruguay. Revista de
Turismo y Patrimonio Cultural, 6(3), 481–492.
Cannonier, C., & Burke, M. G. (2019). The economic growth
impact of tourism in small island developing states – Evi-
dence from theCaribbean.TourismEconomics, 25(1), 85–
108.
Cárdenas-García, P. J., Sánchez-Rivero,M.,&Pulido-Fernán-
dez, J. I. (2015).Does tourismgrowth influence economic
development? Journal of Travel Research, 54(2), 206–221.
Chen, C., & Chiou-wei, S. Z. (2009). Tourism expansion,
tourism uncertainty and economic growth: New evi-
dence from Taiwan and Korea. Tourism Management,
30(6), 812–818.
Cortes-Jimenez, I., & Pulina, M. (2010). Inbound tourism
and long-run economic growth. Current Issues in Tour-
ism, 13(1), 61–74.
Dogru, T., & Bulut, U. (2018). Is tourism an engine for eco-
nomic recovery? Tourism Management, 67, 425–434.
Dogru, T., Isik, C., & Sirakaya-Turk, E. (2019). The balance
of trade and exchange rates: Theory and contemporary
evidence from tourism. TourismManagement, 74, 12–23.
Driscoll, J. C., & Kraay, A. C. (1998). Consistent covariance
matrix estimation with spatially dependent panel data.
Review of Economics and Statistics, 80(4), 549–560.
Du, D., Lew, A. A., &Ng, P. T. (2016). Tourism and economic
growth. Journal of Travel Research, 55(4), 454–464.
32 | Academica Turistica, Year 13, No. 1, June 2020
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Durbarry, R. (2004). Tourism and economic growth: The
case of mauritius. Tourism Economics, 10(4), 389–401.
Eugenio-Martin, J. L., Martín Morales, N., & Scarpa, R.
(2004). Tourism and economic growth in Latin Ameri-
can countries. ssrn Electronic Journal, 26, 1–28.
Fuinhas, J. A., &Marques, A. C. (2012). Energy consumption
and economic growth nexus in Portugal, Italy, Greece,
Spain and Turkey. Energy Economics, 34(2), 511–517.
Garza, N., & Ovalle, M. C. (2019). Tourism and housing
prices in Santa Marta, Colombia: Spatial determinants
and interactions.Habitat International, 87, 36–43.
Gunduz, L., &Hatemi-J, A. (2005). Is the tourism-led growth
hypothesis valid for Turkey? Applied Economics Letters,
12(8), 499–504.
Habib, M. M., Mileva, E., & Stracca, L. (2017). The real ex-
change rate and economic growth: Revisiting the case us-
ing external instruments. Journal of InternationalMoney
and Finance, 73(1), 386–398.
Hatemi-J, A., & Irandoust, M. (2005). Energy consumption
and economic growth in Sweden: A leveraged bootstrap
approach (1965–2000). International Journal of Applied
Econometrics and Quantitative Studies, 2(4), 87–98.
Holzner, M. (2011). Tourism and economic development?
The beach disease? Tourism Management, 32(4), 922–
933.
Husein, J., &Kara, S.M. (2011). Research note: Re-examining
the tourism-led growth hypothesis for Turkey. Tourism
Economics, 17(4), 917–924.
International Monetary Fund. (2018, 16 July). World eco-
nomic outlook update. https://www.imf.org/en/
Publications/WEO/Issues/2018/07/02/world-economic
-outlook-update-july-2018
Irandoust,M. (2019).On the relation between exchange rates
and tourism demand: A nonlinear and asymmetric anal-
ysis. The Journal of Economic Asymmetries, 20, 1–10.
Jouini, T. (2015). Efficient multistep forecast procedures for
multivariate time series. Journal of Forecasting, 34(7),
604–618.
Katircioglu, S. T. (2009). Revisiting the tourism-led-growth
hypothesis for Turkey using the bounds test and Jo-
hansen approach for cointegration. Tourism Manage-
ment, 30(1), 17–20.
Lanza, A., & Pigliaru, F. (2000). Tourism and economic
growth:Does country’s sizematter?Rivista internaziona-
le di scienze economiche e commerciali, 47, 77–85.
Lee, C. C., & Chang, C. P. (2008). Tourism development and
economic growth: A closer look at panels.Tourismman-
agement, 29(1), 180–192.
Malaquias, R., Borges Junior, D., Malaquias, F., & Albertin,
A. (2019). Climate protection or corporate promotion?
Energy companies, development, and sustainability re-
ports in LatinAmerica.Energy Research& Social Science,
54, 150–156.
Malta, G. A. P., Braga, S. S., & Barbosa, M. F. P. (2019). Con-
ceptions of economic development and the understand-
ing of the role of tourism in poverty reduction. Brazilian
Journal of Tourism Research, 13(2), 16–31.
Marques, L.M., Fuinhas, J. A., &Marques, A. C. (2017). Aug-
mented energy-growth nexus: economic, political and
social globalization impacts. Energy Procedia, 136, 97–
101.
Massidda, C., & Mattana, P. (2013). A svecm analysis of
the relationship between international tourism arrivals,
gdp and trade in Italy. Journal of Travel Research, 52(1),
93–105.
Maximiliano, K. (2014). Problems of tourism safety in Latin
America: Foreword two. In P. E. Tarlow (Ed.), Tourism
security: Strategies for effectively managing travel risk and
safety (pp. xxi–xxiv). Butterworth-Heinemann.
Menegaki, A. N., Marques, A. C., & Fuinhas, J. A. (2017). Re-
defining the energy-growth nexus with an index for sus-
tainable economic welfare in Europe. Energy, 141, 1254–
1268.
Nissan, E., Galindo, M. A., & Méndez, M. T. (2011). Rela-
tionship between tourism and economic growth. Service
Industries Journal, 31(10), 1567–1572.
Özokcu, S., &Özdemir, Ö. (2017). Economic growth, energy,
and environmental Kuznets curve. Renewable and Sus-
tainable Energy Reviews, 72, 639–647.
Payne, J. E., &Mervar, A. (2010). Research note: The tourism-
growth nexus in Croatia. Tourism Economics, 16(4),
1089–1094.
Pesaran, M. H. (2007). A simple panel unit root test in the
presence of cross-section dependence. Journal of Applied
Econometrics, 22(2), 265–312.
Risso, W. A., & Brida, J. G. (2008). The contribution of
tourism to economic growth: An empirical analysis for
the case of Chile. European Journal of Tourism Research,
2(2), 178–185.
Rivera,M.A. (2017). The synergies between human develop-
ment, economic growth, and tourism within a develop-
ing country: An empirical model for ecuador. Journal of
Destination Marketing and Management, 6(3), 221–232.
Santiago, R., Fuinhas, J. A., & Marques, A. C. (2020). The
impact of globalization and economic freedom on eco-
nomic growth: The case of the Latin America and Ca-
ribbean countries. Economic Change and Restructuring,
53(1), 61–85.
Academica Turistica, Year 13, No. 1, June 2020 | 33
José Alberto Fuinhas et al. Tourism and Economic Growth Nexus
Shahzad, S. J. H., Shahbaz, M., Ferrer, R., & Kumar, R. R.
(2017). Tourism-led growth hypothesis in the top ten
tourist destinations: New evidence using the quantile-
on-quantile approach. Tourism Management, 60, 223–
232.
Suárez, J. A., Beatón, P. A., Escalona, R. F., & Montero, O. P.
(2012). Energy, environment and development in Cuba.
Renewable and Sustainable Energy Reviews, 16(5), 2724–
2731.
Tang, C. F., & Abosedra, S. (2014). The impacts of tourism,
energy consumption and political instability on eco-
nomic growth in the mena countries. Energy Policy, 68,
458–464.
Tang, C. H., & Shawn, S. (2009). The tourism – Economy
causality in the United States? A sub-industry level ex-
amination. Tourism Management, 30(4), 553–558.
Toman,M. T., & Jemelkova, B. (2003). Energy and economic
development: An assessment of the state of knowledge.
The Energy Journal, 24(4), 93–112.
Tugcu, C. T. (2014). Tourism and economic growth nexus
revisited: A panel causality analysis for the case of the
Mediterranean Region. Tourism Management, 42, 207–
212.
World Tourism Organization. (2015). Annual Report 2015.
https://www2.unwto.org/sites/all/files/pdf/annual
_report_2015_lr.pdf
World Travel Market. (2018, 26 June). México, Brasil e Ar-
gentina São os Países Mais Visitados na América Latina.
https://news.wtm.com/pt/mexico-brasil-e-argentina
-sao-os-paises-mais-visitados-na-america-latina/
World Travel & Tourism Council. (2018). Caribbean recov-
ery. https://www.wttc.org/priorities/crisis-preparedness
/caribbean-recovery/
Zortuk, M. (2009). Economic impact of tourism on Turkey’s
economy: Evidence from co-integration tests. Interna-
tional Research Journal of Finance and Economics, 1(25),
231–239.
34 | Academica Turistica, Year 13, No. 1, June 2020