Content uploaded by Usman Khalid
Author content
All content in this area was uploaded by Usman Khalid on Jun 12, 2019
Content may be subject to copyright.
1
The Effects of Economic and Financial Crises on International
Tourist Flows: A Cross-country Analysis
Usman Khalid*
School of Economics, University of Nottingham Malaysia
Jalan Broga, 43500 Semenyih, Selangor, Malaysia
Email: Usman.Khalid@nottingham.edu.my. Tel.: +6 (03) 8924 8255.
Luke Emeka Okafor
School of Economics, University of Nottingham Malaysia
Jalan Broga, 43500 Semenyih, Selangor, Malaysia
Email: Luke.Okafor@nottingham.edu.my. Tel.: +6 (03) 8725 3716.
Muhammad Shafiullah
School of Economics, University of Nottingham Malaysia
Jalan Broga, 43500 Semenyih, Selangor, Malaysia
Email: Muhammad.Shafiullah@nottingham.edu.my. Tel.: +6 (03) 8725 3719.
Note: Published version of this article is available at
https://journals.sagepub.com/doi/abs/10.1177/0047287519834360
Please cite this article as: Khalid, U., Okafor, L. E., & Shafiullah, M. (2019). The Effects of
Economic and Financial Crises on International Tourist Flows: A Cross-Country Analysis.
Journal of Travel Research. https://doi.org/10.1177/0047287519834360
* Corresponding author.
2
Abstract
This paper investigates the effect of different economic and financial crises, such as inflation
crisis, stock market crash, debt crisis, and banking crisis on international tourism flows using a
panel gravity dataset of 200 countries over the period 1995 to 2010. The results show that the
inflation crisis has a dampening effect on international tourism flows in both the host and origin
countries. The results also show that domestic debt crisis encourages international tourism
arrivals in the host countries, whereas its impact on international tourism services in originating
countries is negative. Further, the impact of these crises on tourism is region dependent. In
particular, banking crisis depresses international tourism flows in host countries situated in
regions such as America and Latin America and Caribbean, whereas its impact on originating
countries located in regions such as Asia and the Middle East is insignificant.
Keywords: Inflation Crisis, Debt Crisis, Banking Crisis, Gravity Panel Data, International
Tourism
3
Introduction
Economic and financial crises often have persistent and devastating macroeconomic impacts. For
instance, they usually result in lower GDP growth and rising unemployment (Reinhart and
Rogoff 2009, 2011, Andersson 2016). Crises originating in one sector of the economy tend to
spread across to different economic sectors. However, the impact crises have on the respective
sectors of an economy is unique (Andersson and Karpestam 2014, Dell'Ariccia, Detragiache, and
Rajan 2008). Tourism is one sector that is often adversely hit by economic and financial crises.
In recent times, the major crises to hit the tourism sector include the Asian Financial
Crisis (AFC) in 1997 and the Global Financial Crisis (GFC) in 2007-2008 (Papatheodorou,
Rosselló, and Xiao 2010). The effect of these crises on the tourism sector is still under
investigation in the literature. Some studies report the crippling effects of crises on tourism
(Dwyer et al. 2006, Papatheodorou, Rosselló, and Xiao 2010, Wang 2009). However, some, such
as Hall (2010), question how crises are conceptualized and measured, such as conflating
economic and financial crises with other events such as terrorist attacks, oil crises, and climate
change. As such, the validity of such findings has been called into question. In contrast, others,
including Sheldon and Dwyer (2010), believe that crises are opportunities for the tourism sector
to shape up and become more competitive.
One possible explanation for the less convincing results in the empirical literature that
relates to crises and tourism is that most studies focus only on a few major financial crises such
as the AFC or GFC without disentangling these major events into different types of crises
(Prideaux 1999, Ritchie, Amaya Molinar, and Frechtling 2010). Reinhart and Rogoff (2011)
distinguish economic and financial crises into different types, namely inflation crises, sovereign
debt crises, banking crises, and stock market crashes. In some cases, a crisis relates only to one
4
of these categories, however, in most cases, it relates to more than one category. For instance, a
sovereign debt crisis is often triggered by a banking crisis, forcing the national government to
take over debts in the banking sector (Reinhart and Rogoff 2011, Velasco 1987). Similarly, a
debt crisis may result in the depreciation of the local currency and spillover into an inflation
crisis as a country facing insolvency tend to engage in expansionary monetary policies to reduce
the debt burden (Lában and Sturzenegger 1994).
Consequently, the impact of an economic or a financial crisis on international tourism
will eventually depend on the nature of the crisis, such as whether it is a stock market crash,
inflation, debt or banking crisis or a mixture of crises. Arguably, some types of economic or
financial crises can have more severe or even stimulating effect on international tourism than
others. The debt crisis, for instance, may result in the devaluation of the currency, making
tourism services cheaper and hence increasing tourism flows. In contrast, an inflation crisis may
make tourism services more expensive, resulting in lower tourism flows. Therefore, it is crucial
to distinguish between various types of crises.
Further, past studies only focus on a certain crisis that is global in nature, such as AFC or
GFC, while completely ignoring the localized crises that may have affected a particular country
or region. An example of such a crisis was the hyperinflation crisis in Zimbabwe in 2008 or a
prolonged sovereign debt crisis in Russia in the 1990s. Besides, many of the existing studies
focus on analyzing the impact of the crisis on tourism from the perspective of either origin or
destination countries or on one country or a region. For example, Li, Blake, and Cooper (2010)
and Page, Song, and Wu (2012) explore the effect of the GFC on China and UK, respectively.
Considering these gaps in the literature, this study explores the effects of different
economic and financial crises, such as inflation crisis, stock market crash, domestic and external
5
sovereign debt crisis and banking crisis on international tourism using a panel gravity dataset of
200 countries from 1995 to 2010. This paper contributes to the literature that relates to crisis and
tourism in a number of ways. First, while previous studies focus on one particular global crisis,
we distinguish between different types of economic and financial crises, such as inflation crisis,
sovereign debt crisis, stock market crash and banking crisis inter alia. Second, by utilizing a
comprehensive crises dataset from Reinhart and Rogoff (2011), this paper captures crises that
have both global and local ramifications. Third, we use a gravity model to capture the effect of
crises on both inbound and outbound tourism demand of the country-pairs. More specifically, we
analyze to what extent different types of crises affect international tourism flows from the
perspective of both the origin and destination countries rather than the extant literature that
focuses on the crisis in either origin or destination countries.
Our results show that the inflation crisis has a depressing effect on inbound and outbound
international tourism. In the presence of inflation crisis, the demand for tourism service and
spending decline as high level of inflation rate lead to the erosion of consumers’ purchasing
power. Further, stock market crash dampens international tourism flows in destination countries,
whereas its impact in the originating countries is not significant. During the period of stock
market turmoil, businesses cut back on travel expenses, thereby reducing demand for high-end
hotel rooms. Moreover, domestic debt crisis leads to an increase in international tourism arrivals
in the destination countries, whereas the demand for international tourism services declines in the
originating countries in the presence of a domestic debt crisis. During the period of the domestic
debt crisis, domestic currency usually loses significant value relative to a basket of international
currencies. The destination country is thus more attractive for international tourists as the overall
costs of international tourism services fall. In contrast, when the domestic currency devalues
6
substantially in the originating countries, potential tourists have less purchasing power, and as a
result, the demand for international tourism declines.
The results also show that the effects of different crises on international tourism are
region dependent. For instance, inflation crisis has a strong negative impact on international
tourism flows in destination countries situated in regions such as America, Latin America and
Caribbean (LATCA), whereas its impact in destination countries located in Europe and Sub-
Saharan Africa (SSA) is statistically insignificant. On the one hand, inflation crisis is
accompanied by a sharp drop in consumer purchasing power, thus, explaining the negative link
between inflation crisis and international tourism. On the other hand, in general, the demand for
international tourism services in Africa is quite low, while European countries are close to one
another in terms of distance. Therefore, low demand and distance could help to explain why
inflation crisis is not a strong deriving factor in these regions.
The rest of the paper is organized as follows. The next section presents a literature review
on the crises tourism nexus. Section 3 discusses the data and methodology used in the paper
followed by the discussion of the results. Section 5 concludes the paper.
7
Literature Review
There is an emerging, albeit short, list of studies that try to understand the effects of crises on the
tourism sector. The early ‘crisis literature’ focus on assessing the effect of the AFC on Asia-
Pacific countries, such as Malaysia and Thailand. The methodologies used in this type of studies
involved a qualitative analysis of secondary data. For instance, Kontogeorgopoulos (1999)
explored the interplay between sustainable development and sustainable tourism in Thailand in
the face of the AFC. Due to the crisis, Thailand was forced to forego the long-term ecological
sustainability of tourism and thus, encouraged rapid growth in tourism to earn much needed
foreign currency.
Prideaux (1999) examined the impact of the Asian Financial Crisis on tourism in East
Asia and concluded that the effects of the crisis were not as harmful as anticipated. This,
according to the author, was evidence that tourism was more resilient than previously thought.
The need for further and more comprehensive analysis of the impacts of crises on tourism was
also stressed. De Sausmarez (2004) explored the Malaysian tourism industry’s response to the
Asian Financial Crisis and crisis management capability. The author identified Malaysia’s most
successful response to the AFC as international marketing campaigns and policies aimed at
boosting domestic tourism as well as encouraging arrivals from new markets.
Anderson (2006) assessed the preparedness of the Australian tourism industry to shocks
and crises, including AFC and the Bali terrorist attack. The analysis concluded that the tourism
industry in Australia was unprepared to handle such shocks and crises and learned little from
these events. Ritchie, Amaya Molinar, and Frechtling (2010) tried to determine the impact of the
economic crises – the GFC and other crises – on tourism in North America. They conducted this
‘backgrounder’ analysis by considering the descriptive statistics on tourism in North American
8
countries. Results showed that tourism in Canada and the USA were substantially affected by the
GFC and the effect is expected to linger. In Mexico, however, the effects of economic crises
initiated by the swine flu pandemic, exchange rates, and the weather conditions were more
prevalent than the GFC.
The literature, then evolved to Computable General Equilibrium (CGE) modeling which
involves quantitative simulation, using sectoral input-output analysis, of the effect of shocks and
crises on the economies studied. However, CGE studies are often limited in numbers due to the
requirement of extensive sectoral-level data. Sufficient data on tourism to conduct quantitative
analysis are usually hard to come by. Blake and Sinclair (2003) evaluated the policy response of
the US tourism industry to the 9/11 terrorist attacks. Their study found targeted sector-specific
subsidies and tax reductions to be the most effective means of crisis management in the tourism
sector.
Dwyer et al. (2006) explored the economic effects of crises in 2003 – the US invasion of
Iraq and Severe Acute Respiratory Syndrome (SARS) – on the Australian tourism industry.
These crises negatively affected both inbound and outbound tourism. However, the net effect
was not as severe as postponement of outbound travel, led to higher allocation towards savings,
domestic tourism or the purchases of other goods and services. Li, Blake, and Cooper (2010)
conducted tests for the effect of the GFC on tourism in China. Results indicate that tourism
expenditure declined in 2008-09 in China because of the GFC.
Later, as tourism data became more abundant, econometric modeling of tourism demand
and forecasting started to become more popular. Wang (2009) assessed the impact of economic
crises and macroeconomic activities on international inbound tourism demand in Taiwan. Using
the autoregressive distributive lag bounds tests, the study found a negative impact of crises on
9
international tourism demand in Taiwan. The author found that crises related to safety – i.e.,
natural disasters, disease outbreaks, and terrorist attacks – had a more significant effect than
those related to the economy. Page, Song, and Wu (2012) estimated the effect of the GFC (which
they dubbed an economic crisis) and the Swine Flu epidemic on inbound tourism in the UK.
Findings revealed that the GFC and Swine Flu epidemic had a significant adverse effect on
tourist arrivals in the UK.
Eugenio-Martin and Campos-Soria (2014) assessed the impact of the GFC on the
expenditure of tourists from the different regions of Europe. Their methodology included
Simultaneous Semi-Ordered Bivariate Probit model as well as GIS and nonparametric analysis.
Results indicated that the cutback of tourism expenditure depended heavily on the tourist place
of origin’s as well as the origin’s GDP and GDP growth. Perles-Ribes et al. (2016) explored the
effect of business cycles and economic crises on the tourism competitiveness in Spain. The
methodology incorporated tests for unit roots in Spain share in the world tourism market.
Breakpoint unit root tests indicated that the market shares are stationary, and the breakpoints
corresponded to major crises. In addition, highly intensive crises were found to have a persistent
negative effect on Spanish tourism competitiveness.
Further, other studies implemented forecasting techniques to simulate the effect of crises
on tourism. Smeral (2009) tried to decipher the effect of macroeconomic variables on outbound
tourism demand for 15 EU countries. Using data on real ‘tourism imports’ and exchange rates,
the author projected outbound travel demand in 2009 and 2010 – the aftermath of the GFC. Two
scenarios were considered: one optimistic and the other pessimistic — tourism decline, albeit to
varying extents, in both scenarios for the 15 EU member states. Later, Smeral (2010) modeled
and forecasted outbound tourism demand in Australia, Canada, the United States, Japan, and the
10
EU-15 countries over the period 2009-2010 – the aftermath of the GFC. The forecasting exercise
showed a decline in aggregate demand for foreign travel in 2009. However, the situation was
uncertain for 2010; whether the tourism sector would face decline or stagnation. Song and Lin
(2010) attempted to forecast inbound and outbound tourism to and from Asia against the
backdrop of contemporary financial and economic crisis (mainly the GFC). The forecasting
exercise revealed that the crisis was expected to have a negative impact on inbound and
outbound Asian tourism, but the effect would dampen by 2010.
The literature also saw the advent of interviews and questionnaire surveys as well as case
studies designed to ascertain the impact of crises on tourism. Okumus, Altinay, and Arasli (2005)
examined the effects of the political and economic crises occurring in 2001 on tourism demand
in Northern Cyprus and Turkey. Their study conducted surveys and interviews to gather
information on the impact of the crisis on hotels and their preparedness in anticipating and
countering the effects of the crisis. The results show that most hotels in Northern Cyprus were
found to be unable to anticipate the crisis and unprepared to handle its aftermath. De Sausmarez
(2007) explored how crisis indicators could be used to minimize the negative impacts of crises
on the tourism sector and thus ensure sustainable tourism development. Their methodology
involved interviewing private and public stakeholders in the Malaysian tourism sector. The paper
proposed a new crisis indicator – travel trade – and argued that it held the key to observing
market trends and minimizing risk to the tourism sector.
Song et al. (2011) conducted a case study on the demand for hotel rooms in Hong Kong
against the backdrop of Global financial and economic crises. Their findings indicated that the
most important factors determining demand are the income of the tourist place of origin and
relative price levels. Boukas and Ziakas (2013) tested the effect of the GFC on tourism in
11
Cyprus. The methodology involves ‘semi-structured’ interviews with the parties involved in the
tourism service industry. Their study identified the following as the main impacts on tourism in
Cyprus: loss of competitiveness, lower number of visitors and decreased revenue, the decline in
quality and increased costs.
Alegre, Mateo, and Pou (2013) examined the Spanish household’s tourism consumption
model in the face of the business cycle during the 2006-2010 period using data from the Spanish
Household Budget Survey. Their results indicated that unemployment and a higher risk of a job
loss substantially reduced tourism “participation and expenditure.” Stylidis and Terzidou (2014)
derived a model to survey local residents’ attitude towards tourism in Kavala, Greece. The
survey, conducted on 317 citizens, revealed their concerns about their direct (personal) benefit
from tourism as well as its impact on the state of the economy.
In view of the above review, only a handful of empirical studies assessed the impact of
crises – economic and financial – on tourism. Those that did suffer from a range of weaknesses.
Generally, the extant literature examined individual countries or a specific group of countries.
Specifically, the early studies relied on descriptive statistics to understand the impact crises had
on tourism. However, these studies are incapable of establishing causation between crises and
tourism. Later CGE modeling studies involved a more thorough quantitative analysis of the
causal links between crises and tourism. However, CGE models ‘simulate’ (a form of projection)
the potential effects of crises on tourism. Thus, CGE studies often show what is the ‘expected
effect’ rather than estimate the ‘actual effect’ using post-crisis data.
The more recent econometric and forecasting studies offered valuable insights into the
issue. However, these studies tended to focus on either inbound or outbound tourism demand
model or estimate them separately. In addition, forecasting studies suffer from the same ‘actual
12
vs. expected’ effect problem faced by CGE modeling papers. Further, the case studies,
interviews, and surveys were very limited in scope as they are conducted on narrow groups of
people as well as specific cities and countries. This restricts their ability to decipher the effect of
crises on tourism and establish causality that is applicable in general. In general, no study to date
has explored the effects of different economic and financial crises, such as inflation crisis, stock
market crash, domestic sovereign debt crisis, external sovereign debt crisis, and banking crisis on
tourism using gravity approach and cross-country data.
Thus, our review of the literature identifies a significant gap in the discussion relating to
the impact of crises on tourism. It is therefore pertinent to adopt a comprehensive and rigorous
approach which analyses, using actual data, how different economic and financial crises
influence tourists’ behavior in both origin and destination countries in various parts of the globe.
13
Data and Methodology
Empirical Strategy
The gravity models have been extensively used to explain international trade flows, migration,
and foreign direct investment. For instance, gravity model is applied to estimate the determinants
of international trade flows (Anderson and Van Wincoop 2003, McCallum 1995, Rose 2000),
determinants of migration (Gil-Pareja, Llorca-Vivero, and Martínez-Serrano 2007, Karemera,
Oguledo, and Davis 2000), determinants of bilateral foreign direct investment (Bergstrand and
Egger 2007, Eichengreen and Tong 2007, Head and Ries 2008) and estimating the openness of a
country (Frankel and Romer 1999, Khalid 2017).
As international tourism is a major component of international trade in services, an
augmented gravity model, therefore, can be used to explore the determinants and the patterns of
international tourism (Khadaroo and Seetanah 2008, Morley, Rosselló, and Santana-Gallego
2014, Santana-Gallego, Ledesma-Rodríguez, and Pérez-Rodríguez 2016). Morley, Rosselló, and
Santana-Gallego (2014) provide a theoretical foundation for using a gravity equation to model
tourism demand based on individual utility theory. Further, numerous papers have used a gravity
equation to empirically model tourism demand (Fourie and Santana-Gallego 2013, Khadaroo and
Seetanah 2008, Santana-Gallego, Ledesma-Rodríguez, and Pérez-Rodríguez 2016, Okafor,
Khalid, and Then 2018, Vietze 2012).
We specify the gravity models for estimating the effects of different crises on tourist
flows in line with extant literature. The general tourism demand model from the literature –
including Smeral (2009, 2010), Wang (2009), Song and Lin (2010), Song et al. (2011), Page et
al. (2012) and Shafiullah, Okafor, and Khalid (2018), inter alia – often expresses the
tourism/travel as a function of income as well as dummies that represent exceptional events such
14
as festivals, sporting events and crises: economic and financial as well as those categorized as
natural disasters. Accordingly, we specify and estimate a reduced-form baseline augmented
panel gravity model in which tourist flows in the destination countries is a function of
population, income and a range of crisis dummies. Equation 1 presents the model for crises in
destination countries and is estimated using the Fixed Effects approach
i
:
(1)
where the ‘ln’ prefix denotes natural logarithm transformation; is a constant; is year effects;
is country-pair fixed effects; is an error term that is i.i.d. distributed across country-pairs
over time and are parameters to be estimated; denotes origin, destination and time
period, is population, is GDP per capita measured at purchasing power parity (PPP),
is a dummy for inflation crisis, is a dummy for stock market crash, and
are dummies for domestic and external sovereign debt crises and is a dummy for banking
crisis.
We also assess the impact of different types of crises on international tourist flows when
the crises take place in the origin countries. This baseline gravity model is specified as Equation
(2) and estimated by the Fixed Effects approach as follows:
(2)
15
Estimating the gravity model using Fixed Effects approach, however, does not allow us
to include traditional gravity variables such as geographical features that do not vary over time.
Therefore, an alternative specification using the OLS approach is used to check the robustness of
the results. The OLS approach allows us to control for gravity variables that do not change over
time. The alternative model for exploring the impact of different types of economic and financial
crises on tourist flows in the destination countries using OLS approach is specified as below:
(3)
Similarly, the alternative model for examining the impact of different kinds of economic
and financial crises on tourist flows in the origin countries using OLS approach is specified as
follows:
(4)
where is country of origin fixed effects, is country of destination fixed effects, are
parameters to be estimated, is distance, is contiguity, is common
official language dummy, is common unofficial language dummy, is former
colony dummy, and is landlocked dummy and is island dummy.
16
Further, in order to check if the results are region dependent, the sample of the study was
divided into four sub-samples by geographical locations of the destination and origin countries.
The sensitivity of the results to geographical locations was tested by re-estimating the Fixed
Effects Specifications. The four different regions used for the classifications are Europe; North
America, Latin American and Caribbean; Asia, Oceania, Middle East, and North Africa; and
Sub-Saharan Africa.
We check the robustness of our results by introducing dynamics in the model. This is
achieved using two different approaches. First, instead of using contemporaneous values of the
crisis variables in the model, we use the first lag and estimate the following model using FE for
the crisis in the destination countries.
(5)
Similarly, for the crisis in the origin countries, the following model is estimated using FE:
(6)
Second, we estimate the model by adding both the contemporaneous and one year lagged
values of the crisis variables. This model will provide information on whether the effects of
different types of crises on international tourism are temporary and short-lived or persists over
17
time. The model for the crisis in the destination is given by the following equation and is
estimated using FE.
(7)
Similarly, for the origin countries, the following model is estimated using FE.
(8)
Overview of Data and Description of Variables
The empirical model is estimated using a panel gravity dataset consisting of 200 countries and
covering the period 1995 to 2010. Countries and time period are chosen based on data
availability. The data on different types of crises were made available by Reinhart and Rogoff
(2011). The data have been widely used in previous studies on crises and their economic impact,
including Agnello and Sousa (2012), Andersson (2016), Andersson and Karpestam (2014),
Furceri and Zdzienicka (2012), inter alia.
The data on variables such as Gross Domestic Product, GDP per capita at PPP, and land
area are collected from the World Bank Development Indicators (World Bank 2017). The data
18
pertaining to the flow of visitors distinguished by the country of origin and destination are
collected from the UNWTO database (World Tourism Organisation 2017). The data on
contiguity, colonial link and language are obtained from the Gravity database from Head, Mayer,
and Ries (2010). A further description of the data and their sources can be found in Table A2 in
the Appendix.
Dependent Variable
Following most previous studies, we use tourist arrivals to capture tourism demand (Gil-Pareja,
Llorca-Vivero, and Martínez-Serrano 2007, Khadaroo and Seetanah 2008, Okafor, Khalid, and
Then 2018). Tourist arrivals are a robust measure of tourist flows as it is relatively easier to
ascertain the number of individuals entering a country. Tourist expenditures are sometimes used
as a measure of tourist flows. However, tourist expenditures are not easily accessible, as they
have to be estimated.
Data on tourism receipts included in the balance of payments tend to be limited due to the
problem of inaccuracy (Sinclair 1998). However, tourist arrivals and tourism receipts are highly
correlated (Neumayer 2004). Given the above submissions, tourist arrival is used as the
dependent variable in the empirical analysis.
Explanatory Variables
The main variables of interest in the present study are the measures of different types of crises as
made available by Reinhart and Rogoff (2011). An inflation crisis is set to 1 when the annual rate
of inflation exceeds 20% in time and zero otherwise. This is in contrast with Reinhart and
Rogoff (2004) that defined inflation crisis using a 12-month inflation cutoff of 40 percent or
higher for the post-World War II period. As argued by Reinhart and Rogoff (2011), the median
19
inflation rate for the period 1914-2009 was about 5 percent. Therefore, similar to Reinhart and
Rogoff (2011), a threshold of 20 percent per annum is used to capture the inflation crisis.
A banking crisis is measured by a dummy variable that is set to 1 when a bank run leads
to bank closure, merger or government takeover or when a bank needs for assistance has a
contagious effect on other banking institutions and zero otherwise. As highlighted by Reinhart
and Rogoff (2011), due to data limitations, banking crises are measured by the events mentioned
above. For instance, the time series on the relative price of bank stocks or financial institutions
relative to the market would be a good candidate for capturing banking crisis, however, the time
series are not easily accessible, particularly for developing countries.
Domestic debt deals with national debts issued under domestic law. In general, domestic
debts are mainly held by residents and also designated in local currency. Domestic debt crises
seldom attract widespread attention as the creditors are often not external agents. Further,
incidences of domestic defaults are in most cases reported in the footnotes. It also attracts less
attention as some domestic defaults that resulted in the conversion of foreign currency into local
currency under duress occurred during the periods of hyperinflations or banking crises or a
combination of the two crises (Reinhart and Rogoff 2011). A domestic debt crisis therefore
takes the value 1 when a country defaults on its domestic debt and zero otherwise. Unlike
domestic debt crises, external debt crises that deal with the inability of sovereign governments to
pay their debts when due attract a lot of attention. External debt defaults could result from
nonpayment or refusal to pay debt obligations or debt restructuring. External debt crisis is
therefore set to 1 when a country defaults on its external debt and zero otherwise.
Further, stock market crashes tend to serve as a signal for macroeconomic depressions.
Empirical evidence suggests that the likelihood of minor depression is 31 percent, while that of
20
major depression is 10 percent following a stock market crash (Barro and Ursúa 2017). In the
presence of a stock market crash, investment and consumption are more likely to decline similar
to the period of macroeconomic depression. The demand for tourism services and investment in
the tourism sector during the period of a stock market crash are more likely to be negatively
impacted. To capture the impact of stock market collapse on international tourism flows, a stock
market crash is measured by a dummy that is set to 1 when a country’s primary stock market
experiences a cumulated multi-year real return of –25 percent or less and zero otherwise.
To find out the overall effect of different crises on international tourism, an overall crisis
indicator destination is set to 1 if a destination experiences any type of economic or financial
crisis covered in this study at time and zero otherwise. Similarly, an overall crisis indicator
origin takes the value 1, if any originating country experiences any kind of economic or financial
crisis covered in this study at time and zero otherwise.
Control Variables
Following the literature on the determinants of tourist flows, we control for variables that
potentially affect international tourism flows (Gil-Pareja, Llorca-Vivero, and Martínez-Serrano
2007, Khadaroo and Seetanah 2008). In this study, tourist flows are measured by the number of
tourists arriving at destination from originating country . In this stance, GDP per capita at
purchasing power parity and population are the best variables that can be used to capture
economic size. Tourists are not only attracted to a destination because of cultural and natural
endowments and/or resources, but also supporting services that go together with business and
leisure, such as restaurants, health resorts, infrastructure services (water supply, waste disposal,
communication, electricity), among others (Gil-Pareja, Llorca-Vivero, and Martínez-Serrano
21
2007). Therefore, the supply tourist services are higher if a country is richer and larger in terms
of population. Similarly, per capita GDP of the originating country is expected to have a positive
effect, as international tourism tends to be a normal good for some individuals, while being a
luxury for a significant fraction of individuals. Further, the larger the population of a country, the
higher the number of tourists (Gil-Pareja, Llorca-Vivero, and Martínez-Serrano 2007).
We also control for the distance between the country of origin and destination. Distance
is used as a proxy to capture the cost of travel between the origin and the destination countries.
This is in line with the notion of the gravity model, which suggests that tourist flows between
country-pair would be higher the closer the distance between them. The closer the distance, the
lower the transportation cost, which in turn would lead to an increase in international tourist
flows and vice-versa. As pointed out by Khadaroo and Seetanah (2008) air travel is the dominant
mode of transportation for international tourism. However, the prevalence of price discrimination
in the sale of air tickets makes it difficult for one to estimate the cost of travel between pairs of
countries.
Similar to extant literature, indicator variables that potentially capture other factors that
either increase or decrease the transaction costs of international tourism are controlled for, such
as sharing a common official or unofficial language, sharing a common border, having colonial
linkage, being an Island and/or a landlocked country (Okafor, Khalid, and Then 2018, Gil-Pareja,
Llorca-Vivero, and Martínez-Serrano 2007). More specifically, cultural and language proximity
plays a crucial role in international tourism. We control for the influence of official language on
international tourism by including a dummy variable that is set to 1 if the country of origin and
destination shares a common official language, and zero otherwiseAs shown by Okafor, Khalid,
and Then (2018) common unofficial language helps to promote international tourism. We,
22
therefore, control for common unofficial language by using a dummy, which takes the value one
if the origin and destination countries share a common unofficial language that is spoken by at
least 9% of the population and zero otherwise.
Further, tourist preferences are more likely to influence international tourism flows. We
capture tourist preferences with contiguity and colonial linkage dummies. Contiguity dummy
takes the value one if the destination and origin countries share a border and 0 otherwise.
Colonial linkages are captured using a dummy variable which is set to 1 if the origin country is a
former colony of the destination country or the destination country is a former colony of the
origin country and zero otherwise.
In addition, we also include an island dummy, which takes the value 1 if the origin
country and destination country are an island and 0 otherwise, and a landlocked dummy, which
is set to 1 if the origin and destination country are landlocked and zero otherwise. Island nations
tend to attract more tourists than landlocked countries. Tourists have the option of traveling by
sea if a nation is an Island which is not a possibility if a country is landlocked. However, the
expected link between Island or landlocked variable and international tourism may not hold if
most tourists travel by air.
Some variables, such as trade ties, weather and climate, and political instability are
excluded from the model. For instance, strong trade ties could develop as a result of the origin
and destination countries sharing membership in a regional trade agreement, which in turn can
lead to the construction of advanced transport networks and greater promotion of tourism
between the two parties, thereby improving international tourism flows. Further, weather and
climate can also influence the destination choice of tourists. In addition, political instability
potentially has an effect on tourism flows. If the likelihood that a government will be ousted or
23
undermined by force or unconstitutional means, the less likely tourists would like to visit such as
a country, just like if the probability of terrorism or violence is high (Chasapopoulos, Den Butter,
and Mihaylov 2014). Given that, our preferred model is a fixed effects specification, the
exclusion of these time-invariant indicators or variables does not introduce any bias in the
parameter estimates.
Summary Statistics
Table 1 presents the summary statistics of the data used in the empirical analysis. Column 1
reports the mean for the whole sample. On average, the probability of occurrence of a financial
or economic crisis is 0.74 and 0.71 per year in the destination and origin countries respectively.
With reference to the destination countries, the most commonly occurring crisis is the stock
market crash which has a probability of occurrence of 0.277 per year, followed by the banking
crisis and external sovereign debt crisis with probabilities of 0.16 and 0.11 per year respectively.
Similar to destination countries, the three leading types of crises are stock market crash, followed
by the banking crisis and the external debt crisis.
(Insert Table1 around here)
On average, destination countries in North America, Latin American and Caribbean
(America & LATCA) countries experienced fewer number of crises than other countries (0.665
per year) whereas, for Sub-Saharan African (SSA) countries, the average is 1.239 per year. On
the contrary, the America & LATCA countries are more likely to experience crises compared to
countries in other regions (0.755 per year) followed by SSA countries with crises probability of
0.747 per year.
Stock market crashes are the most prevalent type of crisis in all four regions, both in
destination and origin countries. Destination countries in Asia, Middle East, North Africa &
24
Oceania regions are more prone to it with a probability of occurrence of 0.335 per year, followed
by destination countries in the SSA. On the other hand, origin countries in the SSA are most
likely to experience a stock market crash with a probability of 0.281 per year, followed by origin
countries in Europe (0.273 per year).
A banking crisis is the second most prevalent type of crisis in all the regions in terms of
both origin and destination countries apart from destination countries in SSA. The probability of
a banking crisis for origin countries varies between 0.161 per year in Asia, the Middle East,
North Africa & Oceania to 0.173 per year in Europe and SSA. Similarly, with respect to
destination countries, the probability of a banking crisis varies between 0.122 per year in SSA to
0.213 in Asia, Middle East, North Africa & Oceania. It is important to highlight the pattern of
occurrence of crises in SSA countries, where an external debt crisis (0.256 per year), and
inflation crisis (0.227 per year) are the most common types of crises behind stock market
crashes.
In general, the likely influence of key determinants of international tourism such as
distance, language, colonial linkage, landlocked dummy, and the Island dummy also differs
across regions. For instance, European countries tend to be closer to each other and are more
likely to share a common border compared to countries in other regions. On the other hand, Sub-
Saharan African countries are more likely to share a common official language than others,
whereas America, Latin American and Caribbean countries are more likely to share a common
unofficial language than their counterparts.
Table 2 reports the test for multicollinearity. The ‘rule of thumb’ in the econometric
literature is that a tolerance level of less than 0.1 or variance inflation factors larger than 10 is a
25
sign of a serious multicollinearity problem. As shown in Table 2, there is no strong evidence
suggesting the presence of multicollinearity problem.
(Insert Table 2 around here)
26
Empirical Results
The link between different types of crises and international tourism in the originating and
destination countries
The parameter estimates obtained using OLS and Fixed Effects (FE) estimators are
presented in Table 3. Columns 1, 2, 4, and 5 report parameter estimates obtained using FE
estimator. Columns 3 and 6 report parameter estimates obtained using OLS estimator. As shown
in columns 1 and 2, in general, crises broadly measured have a negative effect on international
tourism flows in both the originating and destination countries. The parameter estimate from
column 1 suggests that destination countries experiencing different types of crises face a
reduction in international tourism flows by about 4.40% compared to destination countries that
did not experience different types of crises. Similarly, the demand for international tourism
decreases by about 1.30% for originating countries experiencing different kinds of crises
compared to their counterparts with no such crises.
(Insert Table 3 around here)
The finding that crises dampen international tourism is consistent with the notion that
most tourists spend discretionary income on tourism services. Discretionary spending is more
susceptible to economic uncertainty and volatility (Papatheodorou, Rosselló, and Xiao 2010). In
the presence of different types of economic and financial crises, a significant share of individuals
potentially prefers to spend on essentials only and increase savings. A sharp drop in discretionary
spending during the period of crisis tends to depress both outbound and inbound international
tourist flows.
As reported in columns 1 and 4, population and GDP per capita at purchasing power
parity have a significant positive impact on international tourism for both destination and
27
originating countries. This suggests that populous or wealthier destination countries can afford to
supply a more significant share of tourism services, whereas the demand for international
tourism services is larger for populous or more affluent countries. These findings are consistent
with the ones reported by Okafor, Khalid, and Then (2018).
As reported in columns 2, 3, 5 and 6, the effects of economic and financial crises on
international tourism tend to vary by destination and originating countries. We discuss the
effects of different crises on international tourism focusing on the parameter estimates reported
in columns 2 and 5 for three reasons. First, the parameter estimates obtained using FE and OLS
estimators are relatively similar in terms of signs. Second, that estimates obtained using FE are
robust to potential endogeneity issues arising from permanent country-pair specific effects.
Third, OLS is used as a robust check.
As shown in columns 2 and 5, the inflation crisis has a significant negative impact on
international tourist flows in both destination and originating countries. The coefficient from
column 2 indicates that destination countries experiencing inflation crisis face a drop in
international tourist arrivals of about 12.89% compared to their counterparts not experiencing
such a crisis. In contrast, parameter estimate from column 5 suggests that in the presence of
inflation crisis, originating countries demand for international tourism services fall by about
13.15% compared to originating countries not experiencing inflation crisis.
During the period of inflation crisis, the rate of inflation exceeds at least 20%. A sharp
rise in the general price level corresponds to a sharp drop in consumers’ purchasing power. As a
result, the demand for tourism services and spending decline during the period of inflation crisis.
As reported by the World Tourism Organisation (2013), the rate of inflation in the Maldives
peaked at 17% between 2008 and 2009. As a result, the purchasing power of low-skilled
28
employees was eroded. Further, during the 2008 Global Financial Crisis, a sharp drop in tourists’
spending led to a substantial decrease in the service charge which constitutes a major source of
income for low-skilled workers in the Maldives. The income loss resulting from the decline in
demand for tourism services affected around 20 to 30% of about 7,000 Maldivian low-skilled
tourism employees. In general, about 19% of the Maldivian population, consisting mainly of
low-income workers and vulnerable groups were negatively affected in terms of job losses or
income declines in the tourism industry.
As shown in columns 2 and 5, the stock market crash has a significant negative impact on
international tourist flows in the destination countries, whereas its impact in the originating
countries is not statistically significant. A destination country experiencing major multi-year
stock market turmoil faces a 2.66% drop in international tourist arrivals compared to a
destination country not experiencing such stock market downturns. During the period of stock
market crashes, businesses cut back on travel expenses. As a result, the demand for high-end
hotel rooms tends to decline sharply following a stock market crash. As pointed out by Murgoci
et al. (2009), following the 2008 Global Financial Crisis accompanied with the stock market
turmoil, the demand for tourism services from affluent international leisure and business
travelers declined sharply as the world economy slowed down.
The parameter estimates in columns 2 and 5 indicate that the effect of the domestic debt
crisis on international tourism differs for the destination and originating countries. In the
presence of a domestic debt crisis, international tourism arrivals in the destination countries
increase by about 11.74% compared to destination countries experiencing no domestic debt
crisis. In contrast, when an originating country experiences a domestic debt crisis, the demand
29
for international tourism falls by about 6.85% compared to an originating country that does not
experience such a crisis.
In the presence of a domestic debt crisis in a destination country, the domestic currency
usually loses substantial value relative to international currencies. The host country becomes
attractive for international tourists as the overall costs of international tourism services decline.
In contrast, in the presence of a domestic debt crisis in an originating country, the domestic
currency devalues relative to international currencies. Consumers have less purchasing power
and as a result, the demand for international tourism services decrease.
The parameter estimates in columns 2 and 5 suggest that external debt crisis has a
significant positive impact on the demand for international tourism services in the originating
countries, whereas its impact is statistically insignificant in the destination countries. An
originating country experiencing an external debt crisis faces an increase of about 8.33% in
demand for international tourism services compared to an originating country not experiencing
such a crisis. Perhaps, with the advent of the external debt crisis, the demand for tourism services
from business travelers and affluent investors increase sharply as they seek opportunities for
alternative investments abroad. External debt crisis is accompanied by the economic uncertainty
at home and investors would prefer to invest abroad in this kind of situation.
In general, this result suggests that tourism acts as a life jacket for economies facing a
debt crisis. Greece provides a real-life example of the potential interplay between debt crisis and
international tourism. As reported by Helena (2017), amidst the debt crisis, the tourism industry
in Greece accounted for eight out every ten new job openings in 2016. In 2015, figures from the
Bank of Greece suggest that close to 23.5 million tourists visited, yielding about 24% of the
gross domestic product. In 2016, the figure from the country’s tourism confederation indicated
30
that approximately 27.5 million tourists visited. In 2017, approximately 30 million tourists were
expected to visit Greece, which is approximately ten times its population. The booming tourism
industry helps to mitigate the negative impacts of the debt crisis and economic turmoil.
As reported in columns 2 and 5, the banking crisis has a significant negative impact on
international tourism in both the destination and originating countries. International tourist flows
fall by about 5.07% in a destination country experiencing a banking crisis compared to a
destination country not experiencing such a crisis. Similarly, the demand for international
tourism services declines by about 2.50% in an originating country facing a banking crisis
compared to an originating country not experiencing such a crisis.
Banking crisis usually leads to the credit crunch, fall in investor confidence, decrease in
consumer spending and a heightened level of economic uncertainty. In the presence of banking
crisis, consumers decrease discretionary spending. Investment also decreases as firms and
investors have little access to external capital. This suggests that the demand for international
tourism services falls in the originating countries. Similarly, a destination country facing a
banking crisis is less attractive to tourists. For instance, tourists may face difficulty withdrawing
money from ATMs during the period of the banking crisis.
Overall, the standard gravity variables have their expected signs as shown in columns 3
and 6. For instance, distance has a significant negative impact on international tourism in both
the destination and originating countries. Distance could be considered as a proxy for
transportation costs. Higher transportation costs tend to dampen international tourism. Sharing a
border or sharing an official language or sharing an unofficial language or having colonial ties
has a strong positive impact on international tourism in both the destination and originating
31
countries. These results are expected as these factors act to lower the transaction costs of
international tourism.
Further, landlocked destination countries attract less international tourists, whereas the
impact of the landlocked variable on the originating countries is statistically insignificant. As
expected, Island destination countries attract more international tourists, while, island originating
countries demand more international tourism services. The presence of coastline tends to
promote international tourism as it lowers transaction costs, provides an alternative transport
route and presents extraordinary attractions for tourists that are not available in landlocked
destination countries.
The link between different types of crises and international tourism in the originating and
destination countries by regions
Table 4 presents parameter estimates of the link between different crisis events and
international tourism by regions. As shown in columns 2 and 3, the parameter estimates suggest
that inflation crisis has a significant negative impact on international tourist flows in the
destination countries situated in regions, such as America and Latin America and Caribbean
(LATCA). This implies that international tourist flows fall around 27.10% to 34.03% when the
destination countries in these regions are experiencing inflation crisis relative to their
counterparts that are not. There is no strong evidence that inflation crisis has a statistically
significant impact on international tourism in the destination countries located in Europe and
Sub-Saharan Africa (SSA).
(Insert Table 4 around here)
32
Similarly, as reported in columns 5 to 7, in the presence of inflation crisis, the demand for
international tourism services drop by around 9.61% to 14.53% in originating countries situated
in regions such as Europe, America and LATCA, Asia, Middle East, North Africa and Oceania
compared to their counterparts experiencing no inflation crisis. The impact of inflation crisis on
the demand for international tourism services in originating countries situated in SSA is
insignificantly different from zero.
These findings suggest that the impact of the inflation crisis on international tourism is
region dependent. In general, the inflation crisis tends to dampen international tourism in some
regions, while not having any significant impact in some regions. On the one hand, as discussed
earlier, the inflation crisis is associated with a decline in consumer’s purchasing power, thus,
explaining the negative link between inflation crisis and international tourism. On the other hand,
the demand for international tourism services in countries situated in SSA is quite low, and this
might explain why the effect of the inflation crisis is statistically insignificant.
As presented in columns 1 to 8, the impact of the stock market crash on international
tourism is region dependent. As shown in column 2, a destination country in America and
LATCA experiencing a stock market crash faces a drop in international tourist flows of about
4.21% compared to a destination country not experiencing such a crisis. As pointed out earlier, in
the presence of a stock market crash accompanied by economic uncertainty resulting from it,
businesses cut back on travel expenses. International tourism attractiveness of the destination
countries in America and LATCA decreases as a result of economic uncertainty resulting from
the stock market crash.
Further, as reported in columns 3 and 7, in the presence of stock market crash, a
destination or originating country located in Asia, Middle East, North Africa, and Oceania
33
experiences an increase in international tourist flows or the demand for international tourism
services ranging from 2.94% to 5.02% compared to countries in the regions with no such crisis.
Perhaps, a stock market crash in these regions tends to create arbitrage or other business
opportunities for both international and local investors. The potential existence of business
opportunities following a stock market crash might help to explain the finding that a stock
market crash tends to promote international tourism in Asia, Middle East, North Africa, and
Oceania.
The parameter estimates reported in columns 1 to 8 indicate that the effect of the
domestic debt crisis is also region dependent. In the presence of domestic debt crisis in
destination countries situated in America, LATCA, Asia, Middle East, North Africa, and
Oceania, international tourist flows increase by around 20.32% to 21.05% compared to
destination countries not experiencing such crisis. As noted earlier, the domestic debt crisis is
accompanied by the devaluation of the domestic currency, thus, increasing the international
attractiveness of destination countries in these regions. In contrast, when a destination country
situated in SSA, suffers from the domestic debt crisis, international tourist arrivals drop by
around 13.24% compared to their counterparts that are not debt-stricken. In general, SSA
remains a region where tourism competitiveness is the least developed (Crotti and Misrahi 2017).
As a result, a domestic debt crisis accompanied by a low level of international attractiveness of
countries in SSA in general acts to dampen international tourist flows.
Further, originating countries in America and LATCA faces around 20.78% drop in the
demand for international tourism services if they are experiencing a domestic debt crisis
compared to their counterparts that are not. As discussed earlier, in the presence of a domestic
debt crisis, domestic currency depreciates relative to international currencies. As a result, the
34
demand for international tourism services falls as a significant fraction of potential tourists have
less purchasing power.
As reported in columns 1 to 8, the impact of the external debt crisis is region dependent.
International tourist flows fall by about 11.40% in destination countries in America and LATCA
experiencing external debt crisis. Similarly, in the presence of an external debt crisis in SSA,
international tourist flows drop by 12.28% compared to destination countries that are not debt-
stricken. In contrast, an originating country in Europe that is embroiled in external debt crisis
faces an increase in demand for international tourism services by about 26.74% relative to their
counterparts that are not facing such external debt crisis.
In the presence of an external debt crisis, wealthy investors and business travelers in the
rich European originating countries might demand more international tourism services as they
seek for investment opportunities abroad. In contrast, in the presence of external debt crisis
accompanied with a heightened level of economic uncertainty, the international attractiveness of
destination countries in America, LATCA and SSA falls, and thus, international tourist flows
decline.
As shown in columns 1 to 8, the effect of the banking crisis on international tourism is
region dependent. In the presence of a banking crisis, a destination country in America and
LATCA faces a drop in international tourist arrivals of about 4.60%. A destination country in
Asia, Middle East, North Africa, and Oceania attract about 4.69% lower international tourists,
and a destination country in SSA faces a drop in international tourist flows of about 17.14%
compared to their counterparts that are not facing a banking crisis. There is no strong evidence
suggesting that a banking crisis dampens international tourism in destination countries in Europe.
35
Similarly, during the period of the banking crisis, the demand for international tourism
services declines by around 4.11% in originating countries in Europe, whereas originating
countries in the America and LATCA lower demand for international tourism services by about
6.95% relative to their counterparts that are not facing a banking crisis. There is no strong
evidence indicating that a banking crisis in the originating countries in Asia, Middle East, North
Africa, Oceania, and SSA inhibits international tourism.
In the event of a banking crisis, less efficient banks are more likely to collapse, and
interest rates on loans increase sharply as risk premium rises. As a result of the economic
uncertainty resulting from a banking crisis, discretionary spending declines as consumers
decrease spending and investment falls as firms face a higher level of financial constraints. In
general, the banking crisis depresses international flows in countries in different regions as noted
above. The negative impact of a banking crisis on international tourism is, however, more severe
in destination countries in SSA judging from the magnitudes of the parameter estimates. Limited
tourism infrastructure and a low level of tourism competitiveness in destination countries in SSA
relative to destination countries in other regions might help to explain why the negative impact
of the banking crisis on international tourism is more acute in destination countries in SSA.
Robustness Checks
We carry out a number of robustness checks. It may be argued that the dynamic model is
more appropriate compared to the static model. To address this issue, we estimated two types of
dynamic models. Table 5 presents the results of the dynamic models. Overall, the findings are
similar to those reported in Table 3. Columns 1 and 3 present the results when the one-year lag
of the crisis variables are used in lieu of the contemporaneous values. As expected, an inflation
crisis in the destination country would lead to a reduction in international tourist arrivals.
36
Similarly, a stock market crash also significantly reduces international tourism flows to the
destination country. On the other hand, a debt crisis in the destination is positively related to
international tourist flows as in most cases it leads to a devaluation of local currency making the
destination more attractive. Inflation crisis, domestic sovereign debt crises, and banking crisis in
the origin country, on the other hand, leads to significantly lower tourist flows to the destination
country.
(Insert Table 5 around here)
Further, the parameter estimates are qualitatively similar, if the contemporaneous and
lagged values of the crisis variables are controlled for simultaneously. The results are reported in
columns 2 and 4 of Table 5. As expected, inflation crisis in destination and origin countries
significantly reduces tourist flows to destination countries in period t and t-1. Similarly, stock
market crash in origin and destination also reduces tourist flows, however, the effect is only
significant for a crash in destination countries. A domestic sovereign debt crisis in the destination
countries, on the other hand, significantly increases tourism flows to destination countries,
however, this effect is short-lived and disappears after one year.
A banking crisis in the destination country, on the other hand, significantly reduces
tourist flows in the first period, however, after one year, the banking crisis has a significant
positive effect on tourism flows. One possible explanation for this could be the fact that a
banking crisis is likely to trigger institutional reforms, which would, in turn, lead to an
improvement in the quality of domestic economic and political institutions (Andersson 2016).
This, in turn, can improve the attractiveness of the destination country for tourism as tourists feel
safe and secure as the quality of institutions improves.
37
An external debt crisis in the origin countries significantly increases tourist flows to
destination countries in period t and t-1. This result also corroborates our earlier findings as
presented in Table 3. Nonetheless, a banking crisis in the origin country reduces tourist flows,
however, this effect is only significant after one year. Overall, the findings from dynamic models
provide further support to the findings from our main model and suggest that, for most types of
crisis, the impact on tourism is contemporaneous.
We also check if the results are particularly vulnerable to the issue of double counting. It
could be argued that a crisis has a negative effect on income, but this effect is not one to one. For
instance, the impacts of crises on tourism demand are not clear-cut. First, crisis results in a
reduction in tourists’ current income, which in turn, lead to a reduction in expenditure devoted to
tourism activities but not necessarily cancellation of travel altogether. In addition, a reduction in
income would make a typical tourist to be more responsive to price/expense of destinations and
tourism services. Thirdly, in general, the cutback decisions of tourists rely not only on income
but also on other personal considerations such as security, economic confidence, concerns for
socio-economic issues, the environment, etc (Boulding 1991, Hall 2010, Papatheodorou and
Arvanitis 2014).
In view of the potential that each crisis likely has a negative real income effect, if crisis
variables and GDP per capita are included together in the regression, it could lead to the issue of
double counting. To show that our findings are not particularly susceptible to double counting
problem, we employ the procedure of generated regressors proposed by Gomanee, Girma and
Morrissey (2005). This approach consists of constructing a new regressor from the residuals of
an auxiliary regression of the log of GDP per capita on the crisis indicators and use the new
generated regressor in the main regression in lieu of GDP per capita. This approach allows us to
38
identify the effect of changes in GDP per capita on international tourism that are not related to
the occurrence of a crisis.
ii
The parameter estimates of the robustness check are reported in Table 6. Overall, the
conclusions drawn from the new estimates are similar to the ones reported earlier. In fact, after
controlling for the problem of double counting, the coefficients of the indicators of interest have
increased in terms of magnitude, and they are in general more statistically significant relative to
prior parameter estimates, thus, reinforcing the conclusions of the study. For instance, the
inflation crisis, stock market crash, external sovereign debt crisis, and banking crisis in both
destination and origin countries significantly reduces tourism flows to the destination. Domestic
sovereign debt crisis, on the other hand, leads to an increase in tourist flows.
(Insert Table 6 around here)
39
Conclusion
Economic and financial crises present challenges and, occasionally, opportunities for the
tourism industry. With regard to challenges, in the presence of economic and financial crises,
consumers reduce their discretionary spending, which is expected to lower the demand for
international tourism services. In addition, as investor confidence falls, tourism investment
plummets. Compouned by these effects, the tourism sector is likely to face stagnation as well as
job losses. However, the tourism sector could provide opportunities for economies experiencing
some types of crises. During the period of some economic and financial crises, the tourism
industry tends to be resilient and robust, while many sectors tend to be adversely impacted. In
such instances, the tourism sector might help an economy to mitigate the negative effects of
some economic and financial crises.
This study attempts an in-depth analysis of the impact of crises on tourism. We are the
first to explore the effects of different economic and financial crises, namely inflation crisis,
stock market crash, domestic and external debt crisis, and banking crisis on international tourism
using a panel gravity dataset. Our results indicate that some economic and financial crises
adversely affect international tourism, wheras other types of crises increase internaitonal tourism
or have no effect. This is particularly evident when the findings are considered in the context of
different regions of the world. For instance, the results show that an inflation crisis discourages
international tourism in both the destination and origin countries. Inflation crisis erodes
consumers purchasing power as real incomes of potential tourists are substantially reduced.
Consequently, the demand for tourism services decreases as consumers reduce discretionary
spending. A sharp drop in discretionary spending translates into a sharp decline in tourism
spending.
40
Furthermore, our findings indicate that a domestic debt crisis encourages international
tourism in the destination countries, whereas its impact on demand for tourism services is
negative in the origin countries. A devaluation of domestic currency usually accompanies a
domestic debt crisis. The devaluation of domestic currency makes the destination countries more
attractive for tourists, while it lowers the spending power of tourists in the originating countries.
The findings of this study have important policy implications for the tourism sector.
Policymakers should develop macroprudential policy measures to prevent and/or minimize
system-wide risks in the economy in general and the financial sector in particular. For example, a
wide range of system risk indicators could be used to monitor and predict wide systemic risks.
However, in the event that a crisis is irreversible, policymakers should implement policies that
create an enabling environment for the tourism sector to take advantage of any opportunities. An
enabling environment may include, but not be limited to avoiding visa-restrictions and higher
taxes on travel and tourism-related services in the wake of domestic crises, budget shortfalls, etc.
It is essential that tourism thrive because when a host country faces a domestic crisis, it
often becomes more attractive to international tourists due to an increase in their purchasing
power. The tourism sector will, thus, be able to help mitigate the adverse effects of the crisis on
the host economy through tourism receipts and employment generation. If the tourism sector is
able to seize the opportunities left behind due to some domestic crises, the host economy would
quickly correct itself with the least amount of government intervention and the associated market
distortions. Consequently, government policy should be ‘proactive’ instead of ‘reactive’ and
must be formulated with the participation of the relevant stakeholders as well as to be suitable for
the particular situation/region/country.
41
References
Agnello, Luca, and Ricardo M Sousa. 2012. "How do banking crises impact on income inequality?"
Applied Economics Letters 19 (15):1425-1429.
Alegre, Joaquín, Sara Mateo, and Llorenç Pou. 2013. "Tourism participation and expenditure by
Spanish households: The effects of the economic crisis and unemployment." Tourism
Management 39:37-49.
Anderson, Barbara A. 2006. "Crisis management in the Australian tourism industry: Preparedness,
personnel and postscript." Tourism Management 27 (6):1290-1297.
Anderson, James E, and Eric van Wincoop. 2003. "Gravity with gravitas: A solution to the border
puzzle." American Economic Review 93 (1):170-192. doi: 10.1257/000282803321455214.
Andersson, Fredrik NG. 2016. "A Blessing in Disguise? Banking Crises and Institutional Change."
World Development 83:135-147.
Andersson, Fredrik NG, and Peter Karpestam. 2014. "Short and long term growth effects of financial
crises." In Wavelet Applications in Economics and Finance, edited by M. Gallegati and W.
Semmler, 227-248. Springer.
Barro, Robert J, and Ursúa, José F. 2017. “Stock-market crashes and depressions. Research in
Economics”, 71, 384-398.
Bergstrand, J. H., and P. Egger. 2007. "A knowledge-and-physical-capital model of international trade
flows, foreign direct investment, and multinational enterprises." Journal of International
Economics 73 (2):278-308. doi: 10.1016/j.jinteco.2007.03.004.
Blake, Adam, and M Thea Sinclair. 2003. "Tourism crisis management: US response to September 11."
Annals of Tourism Research 30 (4):813-832.
42
Boukas, Nikolaos, and Vassilios Ziakas. 2013. "Impacts of the global economic crisis on Cyprus
tourism and policy responses." International Journal of Tourism Research 15 (4):329-345.
Boulding, Elise. 1991. “States, boundaries and environmental security in global and regional conflicts.”
Interdisciplinary Peace Research 3(2): 78-93.
Chasapopoulos, Panagiotis., Den Butter, Frank. AG. and Mihaylov, Emil. 2014. “Demand for tourism
in Greece: a panel data analysis using the gravity model.” International Journal of Tourism
Policy 5: 173-191.
Crotti, R, and Tiffany Misrahi. 2017. "The Travel & Tourism Competitiveness Report 2017. Paving the
Way for a More Sustainable and Inclusive Future." World Economic Forum: Geneva,
Switzerland.
De Sausmarez, Nicolette. 2004. "Malaysia's response to the Asian financial crisis: Implications for
tourism and sectoral crisis management." Journal of Travel & Tourism Marketing 15 (4):217-
231.
De Sausmarez, Nicolette. 2007. "Crisis management, tourism and sustainability: The role of
indicators." Journal of sustainable tourism 15 (6):700-714.
Dell'Ariccia, Giovanni, Enrica Detragiache, and Raghuram Rajan. 2008. "The real effect of banking
crises." Journal of Financial Intermediation 17 (1):89-112.
Dwyer, Larry, Peter Forsyth, Ray Spurr, and Thiep Van Ho. 2006. "Economic effects of the world
tourism crisis on Australia." Tourism Economics 12 (2):171-186.
Eichengreen, B., and H. Tong. 2007. "Is China's FDI coming at the expense of other countries?"
Journal of the Japanese and International Economies 21 (2):153-172. doi:
10.1016/j.jjie.2006.07.001.
43
Eugenio-Martin, Juan L, and Juan A Campos-Soria. 2014. "Economic crisis and tourism expenditure
cutback decision." Annals of tourism Research 44:53-73.
Fourie, Johan, and María Santana-Gallego. 2013. "Ethnic reunion and cultural affinity." Tourism
Management 36:411-420. doi: 10.1016/j.tourman.2012.10.002.
Frankel, Jeffrey A, and David H Romer. 1999. "Does trade cause growth?" American economic
review 89 (3):379-399.
Furceri, Davide, and Aleksandra Zdzienicka. 2012. "The consequences of banking crises for public
debt." International Finance 15 (3):289-307.
Gil-Pareja, Salvador, Rafael Llorca-Vivero, and José Antonio Martínez-Serrano. 2007. "The impact of
embassies and consulates on tourism." Tourism Management 28 (2):355-360. doi:
10.1016/j.tourman.2006.04.016.
Gomanee, Karuna, Sourafel Girma, and Oliver Morrissey. 2005. “Aid and growth in Sub‐Saharan
Africa: accounting for transmission mechanisms.” Journal of International Development, 17
(8):1055-1075.
Hall, C Michael. 2010. "Crisis events in tourism: subjects of crisis in tourism." Current issues in
Tourism 13 (5):401-417.
Head, K., and J. Ries. 2008. "FDI as an outcome of the market for corporate control: Theory and
evidence." Journal of International Economics 74 (1):2-20. doi: 10.1016/j.jinteco.2007.04.004.
Head, Keith, Thierry Mayer, and John Ries. 2010. "The erosion of colonial trade linkages after
independence." Journal of international Economics 81 (1):1-14.
Helena, S. 2017. "Tourism is our lifejacket: debt-stricken Greece gets record number of visitors".
Retrieved from The Guardian, London, United Kingdom.
44
Karemera, D., V. I. Oguledo, and B. Davis. 2000. "A gravity model analysis of international migration
to North America." Applied Economics 32 (13):1745-1755. doi: 10.1080/000368400421093.
Khadaroo, Jameel, and Boopen Seetanah. 2008. "The role of transport infrastructure in international
tourism development: A gravity model approach." Tourism Management 29 (5):831-840. doi:
https://doi.org/10.1016/j.tourman.2007.09.005.
Khalid, U. 2017. "The effect of trade and political institutions on economic institutions." Journal of
International Trade and Economic Development 26 (1):89-110. doi:
10.1080/09638199.2016.1206142.
Kontogeorgopoulos, Nick. 1999. "Sustainable tourism or sustainable development? Financial crisis,
ecotourism, and the'Amazing Thailand'campaign." Current Issues in Tourism 2 (4):316-332.
Lában, Raul, and Federico Sturzenegger. 1994. "Fiscal conservatism as a response to the debt crisis."
Journal of Development Economics 45 (2):305-324.
Li, ShiNa, Adam Blake, and Chris Cooper. 2010. "China's tourism in a global financial crisis: a
computable general equilibrium approach." Current Issues in Tourism 13 (5):435-453.
McCallum, John 1995. "National Borders Matter: Canada-U.S. Regional Trade Patterns" American
Economic Review 850:615-623.
Morley, Clive, Jaume Rosselló, and Maria Santana-Gallego. 2014. "Gravity models for tourism
demand: theory and use." Annals of Tourism Research 48:1-10. doi:
10.1016/j.annals.2014.05.008.
Murgoci, Cristiana Stefania, Daniela Firoiu, Emilia Ionescu, and Stefan Alexandru Ionescu. 2009.
"Global recession and its effect on tourism and travel industry." Annals of DAAAM &
Proceedings.
45
Neumayer, Eric. 2004. "The impact of political violence on tourism: Dynamic cross-national
estimation." Journal of Conflict Resolution 48 (2):259-281.
Okafor, Luke Emeka, Usman Khalid, and Terence Then. 2018. "Common unofficial language,
development and international tourism." Tourism Management 67:127-138. doi:
https://doi.org/10.1016/j.tourman.2018.01.008.
Okumus, Fevzi, Mehmet Altinay, and Huseyin Arasli. 2005. "The impact of Turkey's economic crisis
of February 2001 on the tourism industry in Northern Cyprus." Tourism Management 26
(1):95-104.
Page, Stephen, Haiyan Song, and Doris Chenguang Wu. 2012. "Assessing the impacts of the global
economic crisis and swine flu on inbound tourism demand in the United Kingdom." Journal of
travel research 51 (2):142-153.
Papatheodorou, Andreas, and Pavlos Arvanitis. 2014. “Tourism and the economic crisis in Greece:
Regional perspectives.” Région et développement 39: 183-203.
Papatheodorou, Andreas, Jaume Rosselló, and Honggen Xiao. 2010. "Global economic crisis and
tourism: Consequences and perspectives." Journal of Travel Research 49 (1):39-45.
Perles-Ribes, Jose Francisco, Ana Belén Ramón-Rodríguez, Antonio Rubia-Serrano, and Luis Moreno-
Izquierdo. 2016. "Economic crisis and tourism competitiveness in Spain: permanent effects or
transitory shocks?" Current Issues in Tourism 19 (12):1210-1234.
Prideaux, Bruce. 1999. "Tourism perspectives of the Asian financial crisis: Lessons for the future."
Current issues in Tourism 2 (4):279-293.
Reinhart, Carmen M, and Kenneth S Rogoff. 2004. “The Modern History of Exchange Rate
Arrangements: A Reinterpretation.” The Quarterly Journal of Economics, 119(1), 1-48.
doi:10.1162/003355304772839515
46
Reinhart, Carmen M, and Kenneth S Rogoff. 2009. "The aftermath of financial crises." American
Economic Review 99 (2):466-72.
Reinhart, Carmen M, and Kenneth S Rogoff. 2011. "From financial crash to debt crisis." American
Economic Review 101 (5):1676-1706.
Ritchie, JR Brent, Carlos Mario Amaya Molinar, and Douglas C Frechtling. 2010. "Impacts of the
world recession and economic crisis on tourism: North America." Journal of Travel Research
49 (1):5-15.
Rose, Andrew K 2000. "One money, one market: Estimating the effect of common currencies on
trade." Economic Policy 30 (30):9-45.
Santana-Gallego, María, Francisco J. Ledesma-Rodríguez, and Jorge V. Pérez-Rodríguez. 2016.
"International trade and tourism flows: An extension of the gravity model." Economic
Modelling 52:1026-1033. doi: 10.1016/j.econmod.2015.10.043.
Sheldon, Pauline, and Larry Dwyer. 2010. "The global financial crisis and tourism: Perspectives of the
academy." Journal of Travel Research 49 (1):3-4.
Shafiullah, Muhammad, Luke Emeka Okafor, and Usman Khalid. (2018). Determinants of
international tourism demand: Evidence from Australian states and territories. Tourism
Economics. Published online September 20. doi: 10.1177/1354816618800642.
Sinclair, M Thea. 1998. "Tourism and economic development: A survey." The journal of development
studies 34 (5):1-51.
Smeral, Egon. 2009. "The impact of the financial and economic crisis on European tourism." Journal
of Travel Research 48 (1):3-13.
Smeral, Egon. 2010. "Impacts of the world recession and economic crisis on tourism: Forecasts and
potential risks." Journal of Travel Research 49 (1):31-38.
47
Song, Haiyan, and Shanshan Lin. 2010. "Impacts of the financial and economic crisis on tourism in
Asia." Journal of Travel Research 49 (1):16-30.
Song, Haiyan, Shanshan Lin, Stephen F Witt, and Xinyan Zhang. 2011. "Impact of financial/economic
crisis on demand for hotel rooms in Hong Kong." Tourism Management 32 (1):172-186.
Stylidis, Dimitrios, and Matina Terzidou. 2014. "Tourism and the economic crisis in Kavala, Greece."
Annals of Tourism Research 44:210-226.
Velasco, Andres. 1987. "Financial crises and balance of payments crises: a simple model of the
southern cone experience." Journal of development Economics 27 (1-2):263-283.
Vietze, Christoph. 2012. "Cultural Effects on Inbound Tourism into the USA: A Gravity Approach."
Tourism Economics 18 (1):121-138. doi: 10.5367/te.2012.0100.
Wang, Yu-Shan. 2009. "The impact of crisis events and macroeconomic activity on Taiwan's
international inbound tourism demand." Tourism Management 30 (1):75-82.
World Bank. 2017. “World Bank Development Indicator”. Washington, USA.
World Tourism Organisation. 2013. “Economic crisis, international tourism decline and its impact on
the poor”. Madrid, Spain: World Tourism Organisation.
World Tourism Organisation. 2017. “UNWTO's Tourism Statistics, CD-ROM”. Madrid, Spain.
48
Table 1: Descriptive Statistics: Average Values
Variables
Whole
Sample
Europe
America
&
LATCA
Asia,
Middle
East, North
Africa &
Oceania
SSA
Log of tourist flows
6.817
7.969
6.196
6.901
6.159
Overall crisis indicator (Destination)
0.747
0.706
0.665
0.668
1.239
Inflation crisis (Destination)
0.084
0.074
0.058
0.049
0.227
Stock market crash (Destination)
0.277
0.268
0.257
0.335
0.288
Sovereign debt crisis domestic (Destination)
0.028
0.017
0.026
0.012
0.098
Sovereign debt crisis external (Destination)
0.115
0.050
0.092
0.033
0.256
Banking crisis (Destination)
0.160
0.147
0.149
0.213
0.122
Ln Population (Destination)
15.747
15.919
15.881
16.471
15.883
Ln GDP per capita, PPP (Destination)
9.161
9.925
9.391
9.198
7.974
Overall crisis indicator (Origin)
0.717
0.690
0.755
0.683
0.747
Inflation crisis (Origin)
0.075
0.067
0.088
0.071
0.083
Stock market crash (Origin)
0.278
0.273
0.272
0.262
0.281
Sovereign debt crisis domestic (Origin)
0.019
0.014
0.038
0.017
0.020
Sovereign debt crisis external (Origin)
0.083
0.055
0.165
0.076
0.103
Banking crisis (Origin)
0.174
0.173
0.168
0.161
0.173
Ln Population (Origin)
15.905
16.042
15.092
15.765
16.043
Ln GDP per capita, PPP (Origin)
9.259
9.542
9.210
9.223
9.064
Ln Distance
8.581
8.053
8.723
8.632
8.593
Contiguity
0.031
0.051
0.024
0.028
0.042
Common Official Language
0.187
0.050
0.285
0.144
0.319
Common Unofficial Language
0.193
0.052
0.296
0.175
0.268
Colonial Ties
0.018
0.032
0.004
0.014
0.015
Landlocked
0.023
0.032
0.006
0.022
0.040
Island
0.064
0.013
0.113
0.062
0.043
Number of Observations
337239
57120
62139
124698
62076
49
Table 2: Tests of Multicollinearity: Variance Inflation Factors (VIF) and Tolerance
Dependent variable: Ln tourist flows
Variable
VIF
1/VIF
Common Official Language
5.31
0.188
Common Unofficial Language
5.10
0.196
Ln GDP per capita, PPP (Origin)
1.52
0.658
External sovereign debt crises (Origin)
1.44
0.697
Ln GDP per capita, PPP (Destination)
1.43
0.700
External sovereign debt crises (Destination)
1.39
0.720
Ln Distance
1.33
0.753
Domestic sovereign debt crises (Destination)
1.31
0.764
Domestic sovereign debt crises (Origin)
1.31
0.765
Contiguity
1.30
0.769
Inflation crises (Origin)
1.28
0.783
Ln Population (Origin)
1.27
0.786
Inflation crises (Destination)
1.26
0.792
Ln Population (Destination)
1.26
0.795
Stock market crash (Destination)
1.17
0.853
Stock market crash (Origin)
1.17
0.854
Colonial Ties
1.13
0.886
Banking crises (Destination)
1.12
0.894
Banking crises (Origin)
1.11
0.904
Island
1.10
0.912
Landlocked
1.07
0.938
Mean VIF
1.64
Note: The ‘rule of thumb’ in the econometric literature is that a variance
inflation factors greater than 10 or a tolerance level less than 0.1 is a sign of a
severe multi-collinearity problem.
50
Table 3: Link between international Tourism and Crises
Variables
FE
FE
OLS
FE
FE
OLS
Crises in Destination
Crises in Origin
(1)
(2)
(3)
(4)
(5)
(6)
Overall crisis indicator
-0.045***
-0.013*
(0.007)
(0.008)
Inflation crisis
-0.138***
-0.211***
-0.141***
-0.107***
(0.024)
(0.040)
(0.026)
(0.034)
Stock market crash
-0.027***
-0.024
-0.001
0.005
(0.008)
(0.014)
(0.008)
(0.014)
Domestic sovereign debt crisis
0.111***
0.199***
-0.071*
-0.080
(0.027)
(0.074)
(0.037)
(0.075)
External sovereign debt crisis
0.021
0.039
0.080***
0.094**
(0.025)
(0.044)
(0.029)
(0.044)
Banking crisis
-0.052***
-0.020
-0.025**
-0.004
(0.012)
(0.016)
(0.013)
(0.014)
Ln Population (Origin)
0.499***
0.296***
0.581***
0.831***
0.486***
0.499***
(0.106)
(0.098)
(0.148)
(0.167)
(0.145)
(0.106)
Ln Population (Destination)
1.098***
1.089***
1.209***
1.042***
1.010***
1.098***
(0.157)
(0.178)
(0.109)
(0.117)
(0.101)
(0.157)
Ln GDP per capita, PPP (Origin)
0.657***
0.587***
0.711***
0.816***
0.810***
0.657***
(0.058)
(0.052)
(0.057)
(0.065)
(0.057)
(0.058)
Ln GDP per capita, PPP
(Destination)
1.087***
1.204***
1.367***
1.415***
1.385***
1.087***
(0.063)
(0.060)
(0.054)
(0.059)
(0.046)
(0.063)
Ln Distance
-1.682***
-1.452***
(0.009)
(0.008)
Contiguity
0.765***
0.837***
(0.040)
(0.037)
Common Official Language
0.572***
0.486***
(0.028)
(0.026)
Common Unofficial Language
0.335***
0.423***
(0.026)
(0.026)
Colonial Ties
1.139***
0.888***
(0.046)
(0.028)
Landlocked
-0.283***
0.068
(0.073)
(0.046)
Island
0.300***
0.065**
(0.030)
(0.027)
Year Effects
Yes
Yes
Yes
Yes
Yes
Yes
Country Effects (Destination & origin)
Origin)
No
No
Yes
No
No
Yes
Country Pair Effects
Yes
Yes
No
Yes
Yes
No
Observations
70683
55711
55711
76048
62911
62911
R-Squared (Within)
0.299
0.289
0.879
0.305
0.335
0.868
Notes: OLS denotes Ordinary Least Square and FE is Fixed-effects estimator. The dependent variable is log of tourist
flows; Ln denotes natural logarithms. The numbers in parentheses in columns are robust standard errors. All
regressions include constant but they are not reported. Significance at the 1%, 5% and 10% levels is indicated by ***,
** and *.
51
Table 4: Link between international Tourism and Crises by Regions
Variables
Europe
America &
LATCA
Asia, Middle
East, North
Africa &
Oceania
SSA
Europe
America &
LATCA
Asia, Middle
East, North
Africa &
Oceania
SSA
Crises in Destination
Crises in Origin
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Inflation crises
0.042
-0.416***
-0.316***
0.037
-0.101**
-0.143***
-0.157***
-0.051
(0.041)
(0.070)
(0.050)
(0.055)
(0.046)
(0.053)
(0.055)
(0.061)
Stock market crash
0.023
-0.043**
0.029**
0.024
-0.004
-0.013
0.049***
0.008
(0.015)
(0.017)
(0.012)
(0.028)
(0.012)
(0.021)
(0.015)
(0.032)
Sovereign debt crises domestic
0.062
0.191***
0.185**
-0.142*
0.040
-0.233**
-0.079
-0.143
(0.047)
(0.063)
(0.075)
(0.085)
(0.073)
(0.091)
(0.065)
(0.089)
Sovereign debt crises external
0.023
-0.121**
0.050
-0.131***
0.237**
0.090
0.035
0.007
(0.067)
(0.059)
(0.083)
(0.031)
(0.095)
(0.064)
(0.057)
(0.044)
Banking crises
-0.029
-0.047*
-0.048**
-0.188***
-0.042**
-0.072**
0.021
-0.051
(0.024)
(0.025)
(0.021)
(0.054)
(0.017)
(0.031)
(0.027)
(0.070)
Other Controls
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Year Effects
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Country Pair Effects
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Observations
16363
10280
23010
6058
31894
10718
15594
4705
R-Squared (Within)
0.146
0.210
0.369
0.478
0.337
0.322
0.371
0.276
Notes: The table reports Fixed-effects estimates. The dependent variable is log of tourist flows; Ln denotes natural logarithms. The numbers in parentheses
in columns are robust standard errors. All regressions include constant but they are not reported. Significance at the 1%, 5% and 10% levels is indicated
by ***, ** and *.
52
Table 5: The dynamic linkages between international Tourism and Crises
FE
FE
FE
FE
VARIABLES
Crises in Destination
Crises in Origin
Inflation crisist
-0.063**
-0.110***
(0.028)
(0.028)
Stock market crasht
-0.025***
-0.004
(0.007)
(0.007)
Domestic sovereign debt crisis t
0.094***
-0.050
(0.028)
(0.035)
External sovereign debt crisis t
0.006
0.055**
(0.018)
(0.023)
Banking crisis t
-0.072***
-0.010
(0.012)
(0.012)
Inflation crisist-1
-0.111***
-0.089***
-0.117***
-0.054**
(0.026)
(0.025)
(0.026)
(0.023)
Stock market crash t-1
-0.050***
-0.032***
-0.009
-0.005
(0.008)
(0.007)
(0.007)
(0.007)
Domestic sovereign debt crisist-1
0.054*
0.034
-0.095**
-0.070*
(0.029)
(0.030)
(0.038)
(0.037)
External sovereign debt crisist-1
0.024
0.003
0.087***
0.067**
(0.023)
(0.020)
(0.030)
(0.027)
Banking crisis t-1
-0.002
0.042***
-0.031**
-0.026**
(0.011)
(0.010)
(0.013)
(0.011)
Ln Population (Origin)
0.604***
0.551***
0.745***
0.796***
(0.099)
(0.107)
(0.162)
(0.171)
Ln Population (Destination)
0.704***
0.963***
1.007***
1.083***
(0.155)
(0.163)
(0.108)
(0.117)
Ln GDP per capita, PPP (Origin)
0.672***
0.654***
0.833***
0.787***
(0.055)
(0.059)
(0.062)
(0.066)
Ln GDP per capita, PPP (Destination)
1.072***
1.100***
1.360***
1.422***
(0.062)
(0.066)
(0.054)
(0.057)
Year Effects
Yes
Yes
Yes
Yes
Country Pair Effects
Yes
Yes
Yes
Yes
Observations
57,799
53,406
64,866
59,885
R-squared
0.289
0.278
0.346
0.334
Notes: FE is Fixed-effects estimator. The dependent variable is log of tourist flows; Ln denotes natural logarithms. The
numbers in parentheses in columns are robust standard errors. All regressions include constant but they are not reported.
Significance at the 1%, 5% and 10% levels is indicated by ***, ** and *.
53
Table 6: Link between international Tourism and Crises using generated regressors for GDP per Capita
FE
FE
VARIABLES
Crises in Destination
Crises in Origin
Inflation crisis t
-1.036***
-0.815***
(0.055)
(0.057)
Stock market crash t
-0.162***
-0.103***
(0.010)
(0.012)
Domestic sovereign debt crisis t
1.105***
0.676***
(0.066)
(0.072)
External sovereign debt crisis t
-1.397***
-0.984***
(0.085)
(0.089)
Banking crisis t
-0.181***
-0.122***
(0.014)
(0.014)
Ln Population (Origin)
0.499***
0.831***
(0.106)
(0.167)
Ln Population (Destination)
1.098***
1.042***
(0.157)
(0.117)
Ln GDP per capita, PPP (Origin)
0.657***
(0.058)
Ln GDP per capita, PPP (Destination)
1.415***
(0.059)
Ln GDP per capita, PPP Residual (Origin)
0.816***
(0.065)
Ln GDP per capita, PPP Residual (Destination)
1.087***
(0.063)
Year Effects
Yes
Yes
Country Pair Effects
Yes
Yes
Observations
55711
62911
R-squared
0.29
0.34
Notes: FE is Fixed-effects estimator. The dependent variable is log of tourist flows; Ln denotes natural
logarithms. The numbers in parentheses in columns are robust standard errors. All regressions include
constant but they are not reported. Significance at the 1%, 5% and 10% levels is indicated by ***, ** and
*.
54
Table A1: Model Selection tests
Test
H0
Statistic
Result
Wald Test
H0: all of the fixed effect intercepts are zero
F = 205.74
Prob > F = 0.00
Reject H0. Result in favor of FE.
Breusch-Pagan LM Test
H0: the variance of the unobserved fixed effects
is zero
= 2.9 ×105
Prob > = 0.00
Reject H0. Result in favor of RE.
Hausman Test
H0: difference in coefficients of FE and RE not
systematic
= 365.45
Prob > = 0.00
Reject H0. Result in favor of FE.
55
Table A2: Definitions and Sources of Model Variables
Variable
Definition
Source
Tourist flows
(TF)
Tourist flows are the dependent variable in the
specified models. Tourist flows are measured by the
number of tourists arriving at destination from
originating country . They are distinguished by the
country of origin and destination.
UNWTO database
(World Tourism
Organization 2017)
Population
(Pop)
This variable controls for the effect of population on
tourist flows. It is one of the variables used to control
for the economic size of the originating and destination
countries. Population data is distinguished by the
country of origin and destination.
World Bank
Development
Indicators (World
Bank 2017)
GDP per
capita at PPP
(GDPpc)
GDP per capita controls for the effect of income on
tourist flows. It is one of the variables used to account
for the influence of economic size in the underlying
relationship. It is measured at purchasing power parity
distinguished by the country of origin and destination.
World Bank
Development
Indicators (World
Bank 2017)
Overall crisis
indicator
(origin)
The overall crisis indicator (origin) accounts for any
economic and financial crisis occurring in the origin
country. It is set to 1 if an origin experiences any type
of economic or financial crisis covered in this study at
time t and zero otherwise.
Author’s own
calculation based on
the Reinhart and
Rogoff (2011) dataset
56
Variable
Definition
Source
Overall crisis
indicator
(destination)
The overall crisis indicator (destination) accounts for
any economic and financial crisis occurring in the
destination country. The variable is set to 1 if a
destination experiences any type of economic or
financial crisis covered in this study at time t and zero
otherwise.
Author’s own
calculation based on
the Reinhart and
Rogoff (2011) dataset
Inflation crisis
(Inf)
This is a dummy (indicator) variable accounting for the
influence of inflation crisis in either the destination or
origin country in the underlying relationship. The
dummy is set to 1 when the annual rate of inflation
exceeds 20% in time t and 0 otherwise.
Reinhart and Rogoff
(2011)
Stock market
crash (SMC)
This dummy variable is used to capture the effect of
stock market crash in either the origin or destination. A
stock market crash is measured by setting dummy to 1
when a country’s primary stock market experiences a
cumulated multi-year real return of –25 percent or less
and 0 otherwise.
Reinhart and Rogoff
(2011)
Sovereign
domestic debt
crisis (SDCD)
This indicator accounts for domestic sovereign debt
crises in either the origin or destination countries. The
variable takes the value 1 when a country defaults on
its domestic debt and 0 otherwise.
Reinhart and Rogoff
(2011)
57
Variable
Definition
Source
Sovereign
external debt
crisis (SDCE)
This is the dummy for external sovereign debt crises in
either the destination or origin countries. The external
debt crisis dummy is set to 1 when a country defaults
on its external debt and 0 otherwise.
Reinhart and Rogoff
(2011)
Banking crisis
(BC)
A banking crisis, in either the origin or destination, is
denoted by the BC dummy. The dummy is set to 1
when a bank run leads to bank closure, merger or
government takeover or when a bank needs for
assistance has a contagious effect on other banking
institutions and zero otherwise.
Reinhart and Rogoff
(2011)
Distance
(DIST)
This variable controls for the effect of distance between
the origin and destination countries on r tourist flows. It
is measured as the number of kilometers between the
destination and origin countries.
Gravity database from
Head, Mayer, and Ries
(2010)
Contiguity
(CONT)
This dummy variable controls for the effect of
contiguity of the origin and destination countries on
tourist flows. The contiguity dummy takes the value of
1 if the destination and origin countries share a border
and 0 otherwise
Gravity database from
Head, Mayer, and Ries
(2010)
58
Variable
Definition
Source
Common
official
language
(COMOFFL)
This variable is an indicator for the effect of a common
official language between the origin and destination
countries on tourist flows. It is set to 1 if the country of
origin and destination shares a common official
language, and 0 otherwise.
Gravity database from
Head, Mayer, and Ries
(2010)
Common
unofficial
language
(COMNOFFL)
This dummy variable controls for the effect of a
common unofficial language on tourist flows between
the destination and origin. This dummy takes the value
of 1 if the origin and destination countries share a
common unofficial language that is spoken by at least
9% of the population and 0 otherwise.
Gravity database from
Head, Mayer, and Ries
(2010)
Former colony
dummy
(FCOL)
The former colony dummy accounts for the effect of
colonial-era ties on tourist flows in the origin and
destination countries. It takes the value of 1 if the
origin country is a former colony of the destination
country or the destination country is a former colony of
the origin country and 0 otherwise.
Gravity database from
Head, Mayer, and Ries
(2010)
Landlocked
(LAND)
The effect of being landlocked on tourist flows is
captured by landlocked dummy. It is set to 1 if the
origin and destination country are landlocked and 0
otherwise.
Gravity database from
Head, Mayer, and Ries
(2010)
59
Variable
Definition
Source
Island (ISLA)
The effect of being an island on tourist flows is
captured by island dummy. This dummy takes the
value of 1 if the origin country and destination country
are an island and 0 otherwise
Gravity database from
Head, Mayer, and Ries
(2010)
60
i
The choice of Fixed Effects estimator is based on the results of statistical tests. The Wald test, Breusch-
Pagan Lagrange multiplier test and Hausman test are performed to decide between Fixed Effects (FE),
Random Effects (RE), and Pooled Ordinary Least Square (POLS) estimator. The results of these tests are
provided in the Table A1 in the Appendix. The Wald test is used to decide between FE and POLS with
the results favoring FE. Breusch-Pagan Lagrange multiplier test is used to decide between RE and POLS
suggesting RE to be superior. Lastly, the Hausman specification test is used to decide between FE and
RE, and the result is in favor of FE. Therefore, based on these tests, the Fixed Effects estimator has been
chosen as the preferred estimator.
ii
The generated residual variables have been recovered from the following estimated regressions:
variable Ln GDP per capita, PPP Residual (Origin) from Ln GDP per capita, PPP (Origin) = 9.89 – 0.83 ×
inflation crisis – 0.12 × stock market crash + 0.92 × Domestic sovereign debt crisis – 1.3 × external
sovereign debt crisis – 0.12 × banking crisis. variable Ln GDP per capita, PPP Residual (Destination)
from Ln GDP per capita, PPP (Destination) = 9.58 – 0.1 ×inflation crisis – 0.01 × stock market crash +
0.24 × Domestic sovereign debt crisis – 0.34 × external sovereign debt crisis + 0.09 × banking crisis. Note
that in the regression for origin and destination, crisis dummies in destination and origin are used
respectively.
