International Remittances as a Driver of Domestic Tourism Expenditure: Evidence from Mexico

Abstract

This article aims to empirically assess the impact of international remittances on domestic tourism expenditure (DTE) in Mexican households. Using Probit and Tobit regressions with instrumental variables and information from the 2016 National Household Income and Expenditure Survey, our findings show that international migrants’ remittances can improve DTE among Mexican households. Controlling for endogeneity of remittances, the results show a statistically significant and positive impact of international remittances on both the probability and the amount of money spent on domestic tourism. This suggests that international remittances provide the necessary resources to alleviate the liquidity constraints that restrict access to tourism activities for recipient households. This research also highlights the need to encourage public policy recommendations aimed at promoting tourism-related development in Mexico.

Keywords

international remittances expenditures tourism instrumental variables Mexico

Introduction

Previous research has proven the importance of the tourism sector as one of the main pillars of economic growth and development in low- and middle-income countries (Pulido-Fernández and Cárdenas-García 2020; Nunkoo et al. 2020; Faber and Gaubert 2019; Jalil, Mahmood, and Idrees 2013). A potential driver of this sector could be related to international remittances, particularly for labor-sending countries that are the main receivers of these resources. Nevertheless, few studies have focused on the use of international remittances as direct generators of activities that improve the well-being of receiving households through expenditure linked to tourism (Mora-Rivera, Cerón-Monroy, and García-Mora 2019).

Worldwide figures show steady growth in the flow of international remittances; in fact, remittances have become the second largest source of external financing for developing countries (World Bank 2017). However, the international health crisis caused by the COVID-19 pandemic will result in unprecedented economic repercussions for remittances. According to the latest World Bank projections, global remittances will suffer a negative impact, estimated at a historic decrease of 20% for 2020. Yet, despite the ongoing economic recession, the World Bank foresees signs of recovery for the receipt of remittances in low- and middle-income countries in 2021, with growth of 5.6% (World Bank 2020). Increases in remittances will be crucial for developing countries since the monetary flows that international migrants send back home are a key factor for economic and social development (S. Lim and Basnet 2017; J. E. Taylor and Castelhano 2016; Adams 2011).

On the other hand, the tourism sector has become vital for global economic growth (Nunkoo et al. 2020; UNWTO 2019). According to estimates by the World Travel and Tourism Council (WTTC), the tourism industry represents nearly 10.4% of the world’s GDP, while close to 320 million jobs around the world are linked to the tourism sector (WTTC 2019). In fact, over the last five years, one out of every five new jobs in the formal economy was created by this sector, making it one of governments’ best allies in generating employment, especially in developing countries (WTTC 2019). As a result of the influence that tourism has on the economy, several developing countries have found that tourism boosts their economic growth and development (Pulido-Fernández and Cárdenas-García 2020; Paramati, Alam, and Chen 2016; Brida, Cortes-Jimenez, and Pulina 2016). Because of the COVID-19 pandemic, however, a decline greater than 50% is expected in international tourist arrivals during 2020. This could translate into a potential worldwide loss of close to $1 trillion in export revenues from tourism, which will put up to 120 million jobs related to this sector at risk (UNWTO 2020). Yet, in spite of the relevance of tourism and remittances worldwide, the literature analyzing the relationship between tourism and different development indicators includes only a handful of studies that focus on the remittances-tourism link (Mora-Rivera, Cerón-Monroy, and García-Mora 2019; Kumar 2014).

To fill this gap and contribute to the debate on the effects of international remittances on the economic development and well-being of labor-sending countries, this research aims to empirically assess the impact of international remittances on domestic tourism expenditure (DTE) in Mexican households. With this study, we will contribute to the literature in two fundamental aspects. On one hand, international remittances can be identified as one of the factors that promote tourism expenditures within a country, an element that has been absent in the vast literature on the main determinants of tourist spending (Park, Woo, and Nicolau 2020; Rashidi and Koo 2016; Marrocu, Paci, and Zara 2015). On the other hand, existing studies that have analyzed the remittances–tourism nexus may have shown biased results since they failed to consider the endogeneity of remittances and the censoring of tourism expenditures. This is important because they represent two methodological problems that have been widely recognized in literature on tourism expenses and remittances impacts (Saadi 2020; Azizi 2018; Alegre, Mateo, and Pou 2013; Adams 2011; Barquet et al. 2011). Therefore, this research is the first to explicitly incorporate both methodological issues in its analysis, ensuring more statistically reliable and robust results, and additionally allows us to confirm or contrast the findings with those of previous studies.

Two main questions guide this research: (1) Are international remittances a fundamental determinant of DTE? and (2) Considering the censored behavior of tourism expenses in developing countries, is it possible to assert that international remittances have a positive impact on the amount Mexican households spend on domestic tourism? The answers to these questions add to the discussion on the effects of international remittances on the development of migrant-sending communities and provide evidence of the potential of such resources in boosting tourism activities when linked with suitable public policies.

This article is relevant for three fundamental reasons. First, Mexico is Latin America’s main receiver of international remittances. In the last two decades, remittances reaching Mexico have increased at an average rate of 10%; this, despite the crisis of 2008–2009 and the dramatically rising deportations of Mexicans from the United States during the past decade. To date, more than half a million immigrants have been deported from the United States under Donald Trump’s administration.¹ Recent estimates by the BBVA indicate that Mexico’s receipt of remittances could drop by 17% in 2020 due to the economic crisis accompanying the COVID-19 pandemic; however, the same study notes that although the short-term economic impact will be more severe than that of the financial recession of 2008–2009, recovery could be faster. Consequently, remittances would regain current levels within a shorter time frame (BBVA Research 2020). The recovery in the flow of these monetary resources will be of great importance since they are the basis for various activities related to production, consumption, and investment in migrant-sending countries, like Mexico (Kumar et al. 2018; J. E. Taylor and Castelhano 2016; Adams 2011; Yang 2011).

Second, the tourism sector is the third source of foreign exchange in Mexico. In 2018, income from tourism represented 8.7% of the GDP (INEGI 2019), nearly double the average of member nations of the Organization for Economic Co-operation and Development (OECD). As for employment generation, tourism produced close to 2.3 million formal jobs in 2018, equivalent to 6% of the nation’s total jobs (INEGI 2019). Moreover, much evidence points to the benefits of this sector as a trigger for economic growth and development in countries which depend largely on these activities, as is the case of Mexico (Pulido-Fernández and Cárdenas-García 2020; Faber and Gaubert 2019; Paramati, Alam, and Chen 2016; Brida, Cortes-Jimenez, and Pulina 2016). Nevertheless, according to the most recent forecasts, the economic recession brought on by the COVID-19 pandemic could cause tourism consumption in Mexico to decrease by approximately 49.3% in 2020 compared to the previous year. This negative impact on the tourism sector translates into a 27% reduction in jobs in the sector, which means that approximately one million people could lose their employment (CICOTUR 2020).

Third, although a significant body of studies has analyzed the main determinants of tourism expenditures (Park, Woo, and Nicolau 2020; Rashidi and Koo 2016; Marrocu, Paci, and Zara 2015; Brida and Scuderi 2013), to our knowledge, no study has considered international remittances as a factor that can determine households’ tourist spending. In the current crisis caused by COVID-19, the identification of mechanisms that will promote the recovery of domestic tourism activities will be indispensable. International remittances could represent one of those mechanisms, since historical evidence shows that the dynamics of remittances have been counter-cyclical; in other words, migrants send more money to their country of origin in times of crisis and adversity (World Bank 2020; Yang 2011). For example, in March of 2020 the reception of remittances in Mexico totaled 4.007 billion dollars, which is 36% more than in March of 2019 (Banxico 2020). Thus, this study would be the first to address in a clear and statistically robust way how international remittances work as a fundamental driver of DTE, a relationship that takes on even more significance in the current scenario of global crisis.

Although international remittances and tourism activities are two of the main sources of foreign exchange and development in Mexico, few studies have addressed this link and its causal effects. For example, works by Mora-Rivera, Cerón-Monroy, and García-Mora (2019) and Cerón and Mora (2014) have focused on this relationship and found that international remittances have positive effects on local tourism consumption. However, since neither paper analyzes the endogeneity of remittances or censorship in tourism expenditures, their findings may be biased. On one hand, scholars have argued that decisions on international remittances, labor supply, and expenditure patterns are made simultaneously (Saadi 2020; Vadean, Randazzo, and Piracha 2018; Azizi 2018; Bettin, Lucchetti, and Zazzaro 2012). This type of simultaneity in expenditures and income decisions means that we must consider the endogeneity of remittances if our purpose is to identify the causal effect of international remittances on DTE (J. E. Taylor and Castelhano 2016; Wooldridge 2010). On the other hand, scholars have emphasized the need to consider tourism expenditures as a zero-censored variable given that a significant proportion of individuals do not carry out this type of spending, especially in low- and middle-income countries (Ngo and Tsui 2020; Alegre, Mateo, and Pou 2013; Barquet et al. 2011). The econometric technique most frequently used to deal with the zero expenditures problem is a Tobit model (Wooldridge 2010; Tobin 1958), whereby the bias associated with the assumption of a linear functional form in the presence of a censored dependent variable can be overcome; thus, this model sets out that covariates determining the probability of spending on domestic tourism also determine the amount of DTE (Alegre, Mateo, and Pou 2013).

To overcome the aforementioned, we propose an innovative two-stage analysis based on information taken from Mexico’s National Household Income and Expenditures Survey 2016. First, we estimate a Probit model with instrumental variables (IVs) to analyze the effects of international remittances on the probability of household spending on domestic tourism; in other words, to examine whether international remittances can be considered a fundamental determinant of DTE. Second, we estimate a Tobit model with IVs that considers the censoring of data used to compute the effect of international remittances on the amount Mexican households spend on domestic tourism. We also estimate Probit and Tobit models without IV as the benchmark models, which provide us with a baseline to contrast our main IV specifications that control for the endogeneity of remittances. To our knowledge, this is the first study that recognizes and integrates both methodological problems into the analysis (endogeneity and censorship), which can lead to econometric results with a higher degree of statistical certainty. Overall, our estimations using an IV approach point toward a statistically significant positive effect of international remittances on DTE.

The rest of the article is organized as follows. The second section contains a literature review on the relationship among migration, remittances, and tourism. The materials and methods of the empirical study are presented in the third section. The results and discussion are examined in the fourth section. Finally, the conclusions and limitations are presented.

Literature Review: The Nexus among Migration, Remittances, and Tourism

The literature review is divided into four sections. The first section examines the circular relationship between tourism and migration. The second section discusses the theoretical mechanisms and empirical evidence that link remittances to tourism, while the third section analyzes the determinants of tourism expenditures. The final section provides an overview of the use of instrumental variables in Probit and Tobit models within the tourism literature.

Tourism and Migration, a Circular Relationship

To study the effects of remittances on domestic tourism, we must first analyze the link between tourism and migratory decisions, as these decisions determine the flows of international remittances to the country of origin, consequently representing income that allows those left behind to participate in domestic tourism activities.

International migration and tourism arrivals are perhaps the two most important social phenomena involving the movement of people in this new era of globalization (UNWTO 2009). The link between them has been highlighted in several economic and tourism studies (Buckley et al. 2020; Song et al. 2020; Dwyer et al. 2014; Feng and Page 2010). Two main concepts can be identified in the migration-tourism nexus, Tourism-led Migration (TLM) and Migration-led Tourism (MLT), which can be understood as a circular relationship (UNWTO 2009). TLM derives from the rapid growth in international tourism. It is motivated by the creation of networks and the acquisition of knowledge that encourages people to move to places with better economic opportunities; hence, tourism boosts migration (Gheasi, Nijkamp, and Rietveld 2011; Haug, Graham, and Mehmetoglu 2007).

Conversely, migratory movements, which have grown significantly worldwide, directly and indirectly cause an intensification in tourist activities in both countries of origin and destination (Marschall 2017a, 2017b; Balli, Balli, and Louis 2016; Etzo, Massidda, and Piras 2014; Feng and Page 2010; UNWTO 2009). Within this second perspective (MLT), most scholars have addressed, on the one hand, the demands for inbound and outbound tourism (Etzo 2016; Massidda, Etzo, and Piras 2014; Seetaram 2012), and on the other, the benefits of international migration as a driving force to improve tourism (Marschall 2017a, 2017b).

As part of the first group of studies, Etzo (2016) in Japan and Massidda, Etzo, and Piras (2014) in Italy examine the impact of migration on inbound tourism and find that the stock of immigrants in both countries is an important determinant of inbound tourism flows. To deepen their analysis, the authors disaggregate the flows by purpose of visit and identify that the impact of migration on inbound tourism goes beyond visiting friends and relatives to include holidays and business. Meanwhile, in the second set of studies, Marschall (2017a, 2017b) implement a qualitative approach to explore how memories of home affect the African transnational migrants’ touristic activities to their origin countries as well as the motivations and perceived significance of these visits in the migrants’ sense of identity and belonging.

International Remittances as a Potential Mechanism to Boost Tourism Expenditures

In addition to the MLT relationship, we can identify a mechanism through which migration can affect and boost tourism activities in the migrant’s country of origin. This potential driver of domestic tourism is the flow of monetary resources sent to family members who are left behind. Given the extraordinary growth of global remittances over the last few years, scholars and policy makers alike have shown increasing interest in conducting studies on possible effects for two main reasons. First is the desire to determine the economic impacts that remittances have had on receiving communities, which have been linked to the development of these territories (J. E. Taylor and Castelhano 2016; Adams 2011). The second reason relates to the steadiness of this flow of resources and its counter-cyclical nature, particularly during economic crises (World Bank 2020; Saadi 2020; Yang 2011; Rapoport and Docquier 2006).

Literature on remittances suggests some theoretical explanations for the expected impact of these monetary resources on expenditure patterns, which can be reflected in DTE. First, international remittances increase the disposable income of receiving households and relax the budget constraints that can modify consumption patterns (E. Taylor 1999; Lucas and Stark 1985). Consequently, remittances allow receiving households to allocate some of this additional income to recreational or leisure activities, including tourism-related activities. In a review of 50 studies on the effects of remittances around the world, Adams (2011) concludes that “all of the studies reviewed here also find that international migration and remittances reduce labour supply and participation, because non-migrants substitute increased income for more leisure” (Adams 2011, p. 825).

In second place, according to some studies, remittance-receiving households become self-sufficient thanks to the increased savings and investments that international remittances permit (Rapoport and Docquier 2006). Self-sufficiency enables participation in extra activities and increases the demand for private medical and educational services and recreational services, which could translate into increased tourism spending (Yang 2011; Cox and Jimenez 1992; Lucas and Stark 1985). Lastly, aside from contributing to household income, international migration and remittances can connect households to novel markets, societies, and cultures. They can prompt families to replace homemade goods with purchased items and to consume more tourist services. They can also modify households’ information set, risk profile, and other preferences in ways that may alter their investment and consumption marginal utilities (Marschall 2017b; J. E. Taylor and Castelhano 2016; Balli, Balli, and Louis 2016; Villagómez 2007; E. Taylor 1999; Stark 1995). For instance, Villagómez (2007) points out that in Mexico’s case, the return of international migrants to their places of origin for local celebrations results in a very significant economic spill, and tourism is one of the most important expenditures.

Despite these theoretical assertions, empirical evidence on the effects of international remittances linked to DTE is still in its incipient stages. Understanding whether impacts are positive, negative, or nonexistent is an empirical task that probably depends on the context and period of analysis. Therefore, our research aims to contribute key elements to ongoing discussions on this topic.

Few studies have analyzed the remittances–tourism nexus, though some of them analyze how both sectors promote economic growth. For instance, Kumar (2014) shows that tourism has a negative effect on growth in the short run but a positive effect in the long run, while remittances have a net positive effect in the short run and a negative effect in the long run. The author’s main conclusion is that tourism is a leading driver of Kenya’s economic growth and that developing remittance infrastructure is extremely vital in boosting the impact of remittances on tourism, in light of accessibility, affordability, and availability of cost-effective financial services. In another study, Kumar, Naidu, and Kumar (2011) analyzed the contributions of open trade, remittance inflows, and expansion in tourism toward improving income in Vanuatu. The authors found that open trade and remittances have a positive effect on long-run economic growth, while tourism has no effect.

Other research, although incipient, has studied how remittances drive tourism activities. Studies by Gullette (2009) and Cohen and Rodriguez (2005) show that remittances may contribute to local development, providing communities with the opportunity to be part of the region’s development. Reyes et al. (2009) found that international family remittances raise households’ consumption, while nostalgia tourism earnings encourage local production. More recent studies evaluate the effect of remittances on tourism consumption in Mexican households (Mora-Rivera, Cerón-Monroy, and García-Mora 2019; Cerón and Mora 2014). For example, Mora-Rivera, Cerón-Monroy, and García-Mora (2019) investigate the impact of remittances (internal and external) on domestic tourism consumption. Using matching methods, the authors show that remittances have a positive impact on tourism spending, and that the probability of household spending on domestic tourism doubles if the household receives international versus domestic remittances. However, these authors fail to control for the endogeneity of remittances and may yield estimates that are biased and potentially misleading.

Studies on the Determinants of Domestic Tourism Expenditures

A vast number of studies have analyzed tourism spending and its main determinants (Park, Woo, and Nicolau 2020; Mudarra-Fernández, Carrillo-Hidalgo, and Pulido-Fernández 2019; Brida and Scuderi 2013). The literature on the determinants of domestic tourism expenditures is highly mixed, context-specific (Brida et al. 2018; Lima, Eusébio, and Varum 2018; L. Wang et al. 2018; Rashidi and Koo 2016), and focused on using micro data (Boto-García, Baños-Pino, and Álvarez 2019; Mitra et al. 2019; Y. Wang and Davidson 2010).

Most of these articles use economic constraints (e.g., assets, financial difficulties, income), sociodemographic data (e.g., age, education, gender, number of household members), and travel-related variables (e.g., accommodation, activities, destination) as regressors in their analyses (Boto-García, Baños-Pino, and Álvarez 2019; Brida and Scuderi 2013).

Yet the tourism literature contains no evidence that international remittances have been incorporated as a potential determinant of tourist spending. Remittances can boost tourism expenditure, especially in countries with a strong migrant tradition like Mexico’s (Mora-Rivera, Cerón-Monroy, and García-Mora 2019). By incorporating remittances as a regressor, tourism expenditure can be analyzed in a more comprehensive way, something that has not yet been explored in previous literature. This suggests that our study is the first to incorporate and evaluate international remittances as one of the factors associated with domestic tourism expenditure and to identify the causal effect of remittances—especially when considering endogeneity.

Probit and Tobit Models with Instrumental Variables in Tourism Literature

Because individuals make many economic decisions simultaneously, endogeneity problems can arise (e.g., reverse causality and simultaneity) when decision makers try to analyze the causal relationship among them (Wooldridge 2010). A technique that is commonly used to solve these methodological drawbacks is the IV approach, previously employed in tourism studies (Wamboye, Nyaronga, and Sergi 2020; Park, Woo, and Nicolau 2020; Gao, Su, and Wang 2019; Koo, Lim, and Dobruszkes 2017). For instance, Wamboye, Nyaronga, and Sergi (2020), Park, Woo, and Nicolau (2020), and Santos and Vieira (2012) used IVs to analyze the determinants of tourism expenditures and concluded that by controlling for endogeneity, the impact of certain independent variables on tourism expenditures could be dramatically increased (as is the case of education) or, conversely, reduced and even canceled altogether (as happens with the length of stay). Gao, Su, and Wang (2019) and C. Lim, Zhu, and Koo (2018) adopted this technique to examine the relationship between urban development and tourism growth. The authors’ main results demonstrate that investment in infrastructure boosts the demand for tourist services.

Within tourism literature, a handful of studies have used IV Probit and IV Tobit models to analyze the links between tourism and diverse subjects (Ngo and Tsui 2020; Ohe 2020; Gunter 2018; Parida, Bhardwaj, and Chowdhury 2017, 2018; Ohe and Kurihara 2013). Among studies that have implemented an IV Probit model, the work of Gunter (2018) should be highlighted. Using data from Airbnb listings in San Francisco, the author looks at the marginal contributions of four criteria that need to be fulfilled to achieve Superhost status. His results reveal that by far the most important criterion is receiving and maintaining excellent guest ratings. In Japan, Ohe (2020) and Ohe and Kurihara (2013) use data from a survey of agricultural cooperatives to investigate the complementarity of the direct economic effects of local farm production (i.e., increases in income and employment) and the indirect economic effects (i.e., the development of tourism). In both studies, the authors implement an IV Probit model and find that cooperatives are more likely to have an increase in their income and employment levels if they (i) have a quality control system, (ii) preserve the local heritage, and (iii) have a close relationship with the accommodation/restaurant sectors.

Scholars have pointed out that studies using IV Tobit models offer results with a higher degree of statistical confidence (Ngo and Tsui 2020; Parida, Bhardwaj, and Chowdhury 2017, 2018). Ngo and Tsui (2020), for example, have used a sample of 11 airports in New Zealand to estimate airports’ efficiency and to identify key determinants. Their main findings suggest that tourism demand has a positive and significant impact on the performance and efficiency of New Zealand’s airports. In India, Parida, Bhardwaj, and Chowdhury (2018) have analyzed the effect of physical development on domestic and foreign tourist arrivals, as well as the impact of government tourism development on tourism revenue; their methods are IV-2SLS and IV Tobit techniques using panel data from 25 Indian states during 1995–2011. Results from the IV Tobit model show that government spending on the tourism industry is positively correlated with a significant rise in tourism revenues.

Materials and Methods

Data and Variables

The database used in this article was Mexico’s National Household Income and Expenditures Survey 2016 (referred to as ENIGH in Spanish).² The aim of the survey is to provide a statistical overview of the behavior of households’ income and expenditures in terms of amount, origin, and distribution. Additionally, it allows us to collect data on the occupational and sociodemographic characteristics of household members (gender, age, schooling, ethnicity, marital status, occupation, etc.), as well as on housing infrastructure and household equipment (materials of walls, flooring, access to basic services, etc.) (INEGI 2017a). The survey’s sampling design guarantees a representative cross section of Mexico at the national, state and urban/rural levels. It is worth noting that national household income and expenditures surveys are quite useful, not only because of their wealth of sociodemographic information but also because of the detailed information on expenditures (for instance, tourism expenditures). These types of surveys therefore constitute a powerful source of data for the analysis of household expenditures and their determinants from a microeconomic perspective (Mora and Arellano 2016; Alegre, Mateo, and Pou 2009, 2013).

Considering Mexico’s Tourism Satellite Account,³ this research identified DTE as expenses by visitors who are Mexican residents (INEGI 2013). Expenditures in this category included car rental, railway and air transport, highway tolls, foreign transport, lodging, tourist packages, accommodation, and tours. According to data from Mexico’s Tourism Satellite Account, the DTE recorded a growth rate of 3.3% between 2017 and 2018 (INEGI 2019). On the other hand, categories specified by the ENIGH comprise the highest proportion of these expenditures and are linked to spending on tourist packages, lodging, and accommodation (INEGI 2017a).

Although this survey was not designed for the study of migration, it allows us to identify income from abroad and to classify it as international remittances, a process that has frequently been used by other studies on the subject (Pardo and Dávila 2017; Mora and Arellano 2016). In this research, we considered the total international remittances received by a household which can be identified through information from the ENIGH (INEGI 2017a). We are also able to identify how, thanks to these resources, families can solve liquidity constraints that might limit them in allocating resources to tourism activities.

Table 1 presents the definitions and metrics for dependent and independent variables used in the econometric models. As shown, independent variables were grouped into three major blocks: demographic, socioeconomic, and regional characteristics. The dummy tourist expenses and tourist expenses variables, which identify whether the household spent on domestic tourism and the total amount of DTE, respectively, are used as outcome variables.

Table 1.

Definitions and Metrics for Dependent and Independent Variables Used in the Probit and Tobit Models.

Dependent Variable
Dummy tourist expenses	The household spent on domestic tourism = 1, otherwise = 0
Tourist expenses	Quarterly tourism expenditure
Individual or household variables
Demographic characteristics
Gender of household head	If household head is male = 1, female = 0
Age of household head	Age of household head (in years)
Squared age of household head	Squared age of household head (in years)
One-person household	Just one person in the household = 1, otherwise = 0
One child in the household^a	Just one child in the household = 1, otherwise = 0
Two or more children in the household	More than one child in the household = 1, otherwise = 0
Socioeconomic characteristics
Completed years of schooling	Years of completed schooling of the household head
Number of income earners	Number of income earners in the household
Number of economic dependents	Number of household members under 12 and over 65 years old
Poor household	The household receives a monetary transfer from a public program = 1, does not receive = 0
Cell phone in the household	A member of the household has a cell phone = 1, otherwise = 0
Housing ownership	If the house is owned = 1, otherwise = 0
Regional variables
Area of residence	The household is located in an urban area = 1; it is located in a rural area = 0
Traditional region	The household is located in Aguascalientes, Colima, Durango, Guanajuato, Jalisco, Michoacán, Nayarit, San Luis Potosí, and Zacatecas = 1, otherwise = 0
Northern region	The household is located in Baja California, Baja California Sur, Coahuila, Chihuahua, Nuevo León, Sinaloa, Sonora or Tamaulipas = 1, otherwise = 0
Central region	The household is located in Morelos, Querétaro, Tlaxcala, Puebla, Hidalgo, Ciudad de México, or Estado de México = 1, otherwise = 0
Southeastern region	The household is located in Campeche, Chiapas, Guerrero, Oaxaca, Quintana Roo, Tabasco, Veracruz, or Yucatán = 1, otherwise = 0

Source: By authors based on ENIGH (National Household Income and Expenditures Survey) 2016.

In this research, all household members younger than 12 years are considered as children.

The demographic characteristics are the age of the household head (continuous variable that measures age in years), a dummy variable that accounts for the gender of the household head (one for male and zero for female), and three additional dummy variables related to the household composition—one-person household, one child in the household, and two or more children in the household. Since previous empirical studies show a significant link between variables related to household size and tourism expenditures, we created three dummy variables to capture this relationship. Implicit in this relationship is the assumption that households with one child are more likely to spend on tourism; however, as the number of children in a household increases, this probability could decrease. Although gender is a variable that does not significantly impact tourism expenditures in some studies (Marrocu, Paci, and Zara 2015; Alegre, Mateo, and Pou 2013), we decided to include the gender of the household head to control unobservable factors that might be correlated with this variable. Since according to Brida and Scuderi (2013), age appears most frequently in empirical studies related to tourism expenditures, we include the age of the household head in our models.

The set of variables that describe the socioeconomic characteristics of the households includes, on one hand, three continuous variables: completed years of schooling, number of income earners, and number of economic dependents. On the other, three dummy variables were incorporated: poor household, housing ownership, and, to proxy household assets, the variable cell phone in the household. Regarding education, it is commonly considered a social class variable, either in the form of a continuous (years of schooling) or categorical variable (degree level), and is included in most empirical investigations (Marrocu, Paci, and Zara 2015; Alegre, Mateo, and Pou 2013; Zheng and Zhang 2013; Bilgic et al. 2008; Alegre and Pou 2004). Our study defines education as the household head’s completed years of schooling.

Household income at a poverty level is also an important determinant of tourism expenditure. In order to make an approximation, a dummy variable was included to identify households that receive a monetary transfer from a public government program, since the vast majority of them are located in the lowest deciles of income distribution. An additional approximation representing the household’s liquidity was introduced by including two continuous variables: the number of income earners and the number of economic dependents. In the first case, the variable measures the number of household members earning current monetary income, while the second variable counts the number of household members under 12 and over 65 years old. Regarding regional characteristics, five dummy variables were included. The first variable attempts to capture the fact that urban residents are more likely than rural residents to spend on tourism. The remaining four regional variables are included to show the geographic differences in international remittance reception.⁴

Descriptive Statistics

Some interesting descriptive statistics that allow us to identify the differences among five contrasting household groups are presented in Table 2. From the 70,038 households available in our sample, 19,962 (28.5%) were households for which tourism expenditure was recorded, while the remaining 71.5% reported no tourism expenditure. These figures allow us to justify the use of a zero-censored dependent variable model, such as the Tobit model used in our research. In addition, 4,107 households received international remittances, and almost 30% of these households spent on tourism (1,212). The proportion of male-headed households between the group of households that received remittances and spent on tourism and those that did not spend on tourism is quite similar (58.2% vs 57.1%). However, the average age of household heads is statistically different: 51.8 years old for the first group and 53.7 years old for the second. In terms of human capital, it is worth pointing out that important differences in the years of completed schooling exist in these two groups: 7.4 years for household heads living in homes that received remittances and spent on tourism compared to only 5.7 years for households that did not spend on tourism, a statistically significant difference.

Table 2.

Mean Values for Selected Variables.

Variables	Total Households	Households Spending on Domestic Tourism	Households Not Spending on Domestic Tourism	Households Receiving International Remittances and Spending on Domestic Tourism	Households Receiving International Remittances and Not Spending on Domestic Tourism	Testp Value
Individual or household variables
Demographic characteristics
Gender of household head	0.722	0.733	0.718	0.582	0.571	0.541
Age of household head	49.19	47.78	49.81	51.80	53.70	0.000***
One-person household	0.107	0.102	0.109	0.082	0.116	0.000***
One child in the household	0.772	0.811	0.755	0.751	0.744	0.509
Two or more children in the household	0.068	0.047	0.078	0.074	0.083	0.259
Socioeconomic characteristics
Total quarterly expenditure	41,222.8	64,458.1	31,077.4	50,168.1	27,870.3	0.000***
Quarterly tourist expenditure	698.8	2,299.1	–	1,755.8	–	–
Quarterly income from remittances	344.0	400.8	319.3	8,284.8	6,779.3	0.003***
Remittances as percentage of total quarterly expenditure				16.51%	24.32%
Completed years of schooling	8.568	10.555	7.700	7.432	5.729	0.000***
Number of income earners	2.452	2.478	2.440	2.911	2.748	0.000***
Number of economic dependents	1.074	0.915	1.143	1.224	1.307	0.047**
Poor household	0.308	0.216	0.348	0.389	0.507	0.000***
Cell phone in the household	0.836	0.914	0.801	0.847	0.760	0.000***
Housing ownership	0.689	0.684	0.691	0.711	0.731	0.817
Regional variables
Area of residence	0.784	0.850	0.755	0.621	0.530	0.000***
Traditional region	0.224	0.247	0.213	0.517	0.480	0.023**
Northern region	0.220	0.233	0.214	0.164	0.155	0.747
Central region	0.320	0.305	0.327	0.120	0.133	0.225
Southeastern region	0.227	0.209	0.235	0.194	0.225	0.129
Observations	70,038	19,962	50,076	1,212	2,895

Source: By authors based on ENIGH (National Household Income and Expenditures Survey) 2016.

Data in Table 2 also shows that the percentage of poor households that received international remittances and spent on tourism (38.9%) is considerably lower than the percentage of poor households that received international remittances and did not spend on tourism (50.7%). As for socioeconomic characteristics, total quarterly expenditure for households that received remittances and spent on tourism is 50,168 pesos yet only 27,870 pesos for those that did not spend on tourism. Quarterly income from remittances for the first group is 8,248 pesos and 6,779 pesos for the second group. Yet remittances make up 16.5% of the total quarterly expenditure for households that received international remittances and spent on tourism versus 24.3% for those that did not spend on tourism. The above figures suggest that remittances are likely to help boost households’ spending on leisure activities, such as tourism. Finally, households’ demographic and socioeconomic characteristics also include: one-person household (averaging 10.7%); one child in the household (77.2%); number of income earners (2.4); number of economic dependents (1.07); and cell phone in the household (83.6%). These figures are for the whole sample.

Empirical Approach

To identify if international remittances can be considered a main driver of DTE, our empirical approach estimated two econometric models: Probit and Tobit. The former can help determine whether international remittances are a decisive factor when making DTE; in other words, it can tell us whether the probability of a household’s spending on domestic tourism increases or not after receiving international remittances. The latter assesses the effect that variations in the amount of remittances received has on DTE when considering data censoring. Importantly, a crucial element when estimating both models is the potential endogeneity of remittances, which has been pointed out in previous studies (Azizi 2018; Vadean, Randazzo, and Piracha 2018; J. E. Taylor and Castelhano 2016; Bettin, Lucchetti, and Zazzaro 2012). To solve this, we used the instrumental variables technique to overcome the identification issue present in our analysis (Wooldridge 2010; Angrist and Krueger 2001; Heckman 1997).

Probit Model

The Probit model follows the discrete choice models, which link the choice made by each unit of analysis to the attributes of said unit and the attributes of the alternatives available to the unit of analysis. These types of models are often used to forecast how the choices of the unit of analysis (households in this paper) will be altered by changes in demographics and/or attributes of the alternatives. Thus, by using a Probit model we can estimate the probability that a household spends on domestic tourism given a set of explanatory variables.

Following Wooldridge (2010) and Johnston and DiNardo (1997), we assume that the decision of the $i$ th household to spend or not on domestic tourism depends on an unobservable latent variable $y_{i}^{*}$ determined by a set of socioeconomic and sociodemographic characteristics $X_{i}$ , and by the amount of international remittances ( $r e m_{i}$ ) received, in such a way that the larger the value of the latent variable $y_{i}^{*}$ , the greater the probability of a household’s spending on domestic tourism. The latent variable $y_{i}^{*}$ can be expressed as

y_{i}^{*} = α + β r e m_{i} + X_{i} δ + u_{i}^{1},

(1)

where $r e m_{i}$ is the amount of international remittances received by the $i$ th household.

Let $y = 1$ if the household spends on domestic tourism and $y = 0$ if it does not. We can assume that the latent variable has a threshold level, call it $τ$ , such that if $y_{i}^{*}$ exceeds $τ$ , the household will spend on domestic tourism; otherwise, it will not. As with $y_{i}^{*}$ , $τ$ is not observable, but if we assume that it is normally distributed with the same mean and variance, then it is possible to estimate the parameters of the latent variable given in equation (1) and to obtain information about $y_{i}^{*}$ .

Given the assumption of normality, the probability that $τ$ is less than or equal to $y_{i}^{*}$ can be computed from the standardized normal cumulative distribution function (CDF) as (Wooldridge 2010):

\begin{array}{l} P_{i} = P (y = 1 | r e m_{i}, X_{i}) = P (τ \leq y_{i}^{*}) \\ = P (Z_{i} \leq α + β r e m_{i} + X_{i} δ) = F (α + β r e m_{i} + X_{i} δ), \end{array}

(2)

where $P (y = 1 | r e m_{i}, X_{i})$ means the probability that a household spends on domestic tourism given the values of $r e m_{i}$ and $X_{i}$ , and where $Z_{i}$ is the standard normal variable, that is, $Z ~ N (0, σ^{2})$ . $F$ is the standard normal CDF, which written explicitly is

F (y_{i}^{*}) = \frac{1}{\sqrt{2 π}} \int_{- \infty}^{I_{i}} e^{- z^{2} / 2} d z = \frac{1}{\sqrt{2 π}} \int_{- \infty}^{α + β r e m_{i} + X_{i} δ} e^{- z^{2} / 2} d z .

(3)

Now, to obtain information on $y_{i}^{*}$ (the latent variable) as well as on $α$ , $β,$ and $δ$ , we take the inverse of equation (2) to obtain

y_{i}^{*} = F^{- 1} (y_{i}^{*}) = F^{- 1} (P_{i}) = α + β r e m_{i} + X_{i} δ,

(4)

where $F^{- 1}$ is the inverse of the normal CDF.

Since the model is nonlinear, the parameters do not necessarily represent the marginal effect of the independent variables. Thus, to compute the effect of a unit change on one of the independent variables—for example $r e m_{i}$ (international remittances)—on the probability that $y = 1$ , that is, that a household spends on domestic tourism, we can take the derivative of equation (2) with respect to $r e m_{i}$ . This derivative can be expressed as follows (Wooldridge 2010):

\frac{\partial P_{i}}{\partial r e m_{i}} = f (α + β r e m_{i} + X_{i} δ) β,

(5)

where $f (α + β r e m_{i} + X_{i} δ)$ is the standard normal probability density function assessed at $α + β r e m_{i} + X_{i} δ$ . It is worth noting that this assessment will depend on the particular values of $X_{i}$ and $r e m_{i}$ .

Tobit model

In addition to analyzing how international remittances may affect household participation in domestic tourism, we want to examine their impact on DTE, in absolute terms. To attain this objective, we implement a Tobit model which considers the censoring of data used. The Tobit model is an extension of the Probit model (Amemiya 1984). With the Tobit model our interest is to find out the amount of money a household spends on domestic tourism in relation to socioeconomic and sociodemographic variables, and specifically in relation to the amount of international remittances received.

Since only a small percentage of Mexican families spend on domestic tourism, the variable that measures DTE takes the value of zero for a significant proportion of households included in the sample, meaning that we have a censored sample. A model which takes this into account is called a censored regression model. Considering that DTE has roughly a continuous distribution over positive values, we follow Wooldridge (2010) and estimate the following Tobit model:

\begin{array}{l} y_{i} = \max (0, y_{i}^{*}), \\ y_{i}^{*} = α + β r e m_{i} + X_{i} δ + u_{i}^{2} u_{i}^{2} | r e m_{i}, X_{i} \\ ~ N o r m a l (0, σ^{2}), \end{array}

(6)

where $y_{i}^{*}$ is a latent variable; $y_{i}$ is observed domestic tourism expenditure for the $i$ th household; $α$ , $β$ , and $δ$ are parameters to be estimated; and $u_{i}^{2}$ is a normally distributed error term.

Probit and Tobit Models with Instrumental Variables

The ideal scenario for analyzing the impact of international remittances on some key development variables in places of origin would be a situation where remittances would flow and be assigned randomly among receiving households (McKenzie and Yang 2010). This would enable us to identify the causal effect of remittances on relevant outcome variables (poverty, expenditure patterns, investment, education, etc.) (Ambler, Aycinena, and Yang 2015; Adams 2011). Unfortunately, no surveys or information are available in Mexico to conduct an empirical exercise of this nature.

Identifying the effect of international remittances on tourism expenditure is not incompatible with the above description. Given that in the specifications of our models unobserved heterogeneity could explain, at least partially, both the amount of remittances received by a household and its decision to spend on tourist services, it is plausible to consider the possibility of finding an endogeneity issue in our estimations. Consequently, before building any arguments on the causal effects of remittances on DTE, we must implement a technique that can solve this problem.

Because of the impossibility of randomizing the allocation of international remittances among households in this research, as well as the lack of evidence of an exogenous shock that might lead to the development of a natural experiment, the best option to deal with the endogeneity problem was using the instrumental variables method.⁵ To implement this technique, the first step, and perhaps the main one, is finding a valid instrument, that is, one that correlates to the endogenous explanatory variable but not to the error term in the main equation. According to Angrist and Krueger (2001), having a precise understanding of the economic mechanisms that determine variables of interest can lead to good instruments.

In literature, a wide range of variables has been used as instrument for remittances, depending on the data available and the aims of the analysis. These instruments include (1) economic incentives, such as salary gaps, local currency depreciation rates, and employment rate differentials between place of origin and destination (scholars argue that these types of variables influence the amounts of remittances sent but not how these are spent by households in communities of origin) (Acosta et al. 2008; Yang 2008; McKenzie and Rapoport 2007); (2) rainfall patterns, which can be useful for analyzing outcomes in the receiving country (Lopez-Cordova 2006; Munshi 2003); (3) distance between places that send and receive remittances (an instrument that considers that a correlation between distance and the probability of migrating exist, but does not correlate to the potential earnings that a migrant might make in the country of destination nor to the amounts and frequency of remittances sent home) (López-Feldman and Chávez 2017; McKenzie, Stillman, and Gibson 2010; Adams and Page 2005); and (4) historical events, which have significantly shaped migratory patterns and often determine the evolution of migratory networks (which have prompted migratory flows and decisions to send remittances, although with no added impact on current income and expenditure in communities of origin) (Berloffa and Giunti 2019; Woodruff and Zenteno 2007).

In this research, we adopted a combination of the two latter approaches because we consider that transport costs (distance to the US border) together with historical events (the Bracero Program) can be used to identify remittance flows to Mexico. Scholars have shown that in Mexico’s case, the migratory patterns between Mexico and the United States can efficiently solve endogeneity issues linked to remittances (López-Feldman and Chávez 2017; Demirgüç-Kunt et al. 2011; Woodruff and Zenteno 2007).

Thus, the instrument we use is based on the correlation between current migratory routes and the railway network that existed in Mexico in 1920. More specifically, the instrumental variable used is a proxy variable which identifies the relationship between migratory patterns and remittance transfers. This variable is the weighted sum of the distance between the train station closest to the municipal seat in the community of origin plus the distance from said station to the US border ( $s u m_w e i g h t_d i s t$ ). As Woodruff and Zenteno (2007) point out, the literature documenting the geographic origins of migratory patterns between Mexico and the United States shows that this type of instrumental variable is both relevant and exogenous.

A historical event that played a key role in defining current migratory patterns was the Mexican Farm Labor Program (Bracero Program). This program was promoted from 1942 to 1964 in response to US demand for human labor during and after the Second World War and consisted of transporting experienced Mexican farmworkers to the United States. For twenty-two years, the Mexican Farm Labor Program sponsored the crossing of close to 4.5 million guest workers from Mexico (Durand 2016). Most of these workers arrived to the United States by train, traveling on the train tracks built in Mexico between 1884 and 1900. This mass out-migration set the basis for the migratory networks that remain in place even today (Woodruff and Zenteno 2007). Based on this, we can conclude that the link between the railway and migratory networks can explain the correlation between early and current migrations.

Therefore, to solve the potential endogeneity of international remittances in the Probit and Tobit models, a reduced version of the equation was estimated to link the amount of international remittances and the instrument:

r e m_{i} = ω + φ s u m_w e i g h t_d i s t_{i} + X_{i} γ + v_{i},

(7)

where $s u m_w e i g h t_d i s t_{i}$ is the instrumental variable; $ω$ , $φ,$ and $γ$ are parameters to be estimated; and $v_{i}$ is an error term assumed to be normally distributed. Under the assumption that $(u_{i}^{1}, v_{i})$ are jointly distributed as a multinomial normal variable, the model represented by equations (1) and (7) can be estimated as a Probit with instrumental variables. Similarly, under the assumption of the joint normality of $(u_{i}^{2}, v_{i})$ , we can estimate a Tobit model with instrumental variables for DTE (Cameron and Trivedi 2010).

Results and Discussion

Validity and Reliability of IV Technique

Having presented the empirical justification for the use of the variable $s u m_w e i g h t_d i s t$ as a valid instrument to estimate equations 1 and 6, and before discussing results in detail, we will review some tests for the validity of our instrumental variable. Given that it is not possible to calculate the proper tests of the IV Probit and IV Tobit models in a straightforward manner, we present and discuss indicative tests obtained from a standard IV/2SLS model, under the assumption that if these tests suggest the (in)validity of the IV then we can expect the same (in)validity in the correctly specified model. Table 3 shows the results of the tests; the IV is significant in the explanation of the reception of remittances and the underidentification tests are clearly passed. The hypothesis of weak instrument is clearly rejected for international remittances. The final test in Table 3 rejects the exogeneity of international remittances. Considering this information as a whole, we conclude that the use of IV technique is necessary, and that the proposed IV is valid.

Table 3.

Tourism and Remittances. Indicative Instrumental Variable Tests.

	(1)
	International Remittances
Instrumental variable	sum_weight_dist
Underidentification test (F test of excluded instruments)	79.47
H₀: instruments are jointly irrelevant in first stage	0.000
Underidentification test (Kleibergen-Paap rk LM statistic)	78.95
H₀: model is underidentified, instruments are not good	0.000
Weak identification test (Cragg-Donald Wald F statistic)	50.31
H₀: weakly identified system	—^a
Endogeneity test of endogenous regressors	43.26
H₀: variables can be considered as exogenous	0.000

Stock-Yogo weak ID test critical values: 10% maximal IV size 16.38, 15% maximal IV size 8.96, 20% maximal IV size 6.66, 25% maximal IV size 5.53.

Source: By authors based on equation (7).

Econometric Results

To facilitate the interpretation of the impacts of our independent variables on the probability or amount of spending on DTE, marginal effects at the means (MEM) were calculated. To interpret the results, it is worth considering that, in the case of binary variables (e.g., gender), MEM measure the discrete change, while for continuous variables, MEM measure the instantaneous rate of change (Wooldridge 2010).⁶

IV Probit Results

Table 4 shows the results of the estimations analyzing the relationship between the amount of international remittances received and the probability of household spending on domestic tourism.⁷ Altogether, these results are quite promising; however, if we focus on the coefficient results for international remittances, important differences can be observed between specifications that use IV and those that do not. Column (1) shows the results of a Probit model that includes only one explanatory variable: the amount of international remittances received by a household. Although the effect found is positive and statistically different from zero, this model does not consider the endogeneity of remittances and therefore the estimated coefficient could be biased. Column (2) shows the model estimated using the IV (sum_weight_dist). Here, the effect found is larger than in the previous case, and it continues to be positive and statistically significant at 1%, indicating that an increase in international remittances is associated with a higher probability of household spending on domestic tourism.

Table 4.

Probability of Participation in Domestic Tourism Expenditure (DTE).

Dependent Variable: Dummy for DTE
	(1)	(2)	(3)	(4)	(5)	(6)
International remittances	0.0000017***	0.00006***	0.0000032***	0.00013***	0.0000033***	0.00017***
International remittances	[3.26]	[2.95]	[5.86]	[6.45]	[5.86]	[6.01]
Gender of household head			−0.0035	−0.0061	−0.00063	−0.0034
Gender of household head			[−0.87]	[−1.53]	[−0.16]	[−0.86]
Age of household head			−0.0030***	−0.0029***	−0.0031***	−0.0031***
Age of household head			[−4.46]	[−4.35]	[−4.59]	[−4.49]
Squared age of household head			0.00004***	0.00004***	0.00004***	0.00004***
Squared age of household head			[6.57]	[6.55]	[6.48]	[6.40]
Completed years of schooling			0.0234***	0.0235***	0.0226***	0.0226***
Completed years of schooling			[55.00]	[55.11]	[51.47]	[51.51]
One-person household			0.0022	0.0008	0.0023	0.0008
One-person household			[0.35]	[0.13]	[0.36]	[0.13]
One child in the household			0.0292***	0.0287***	0.0281***	0.0275***
One child in the household			[4.35]	[4.26]	[4.17]	[4.08]
Two or more children in the household			−0.0226***	−0.0229***	−0.0216**	−0.0218**
Two or more children in the household			[−2.59]	[−2.63]	[−2.47]	[−2.50]
Number of income earners			0.0331***	0.0328***	0.0319***	0.0317***
Number of income earners			[21.72]	[21.57]	[20.87]	[20.77]
Number of economic dependents			−0.0185***	−0.0180***	−0.0188***	−0.0185***
Number of economic dependents			[−5.68]	[−5.53]	[−5.78]	[−5.69]
Poor household			−0.0554***	−0.0567***	−0.0491***	−0.0480***
Poor household			[−12.48]	[−12.75]	[−10.76]	[−10.52]
Housing ownership			0.0074*	0.0070*	0.0115***	0.0123***
Housing ownership			[1.85]	[1.75]	[2.87]	[3.07]
Cell phone in the household			0.0828***	0.0829***	0.0808***	0.0792***
Cell phone in the household			[17.17]	[17.20]	[16.61]	[16.21]
Area of residence					0.0373***	0.0373***
Area of residence					[9.53]	[9.54]
Northern region					−0.0271***	−0.0289***
Northern region					[−6.29]	[−6.73]
Central region					−0.0363***	−0.0392***
Central region					[−7.41]	[−8.04]
Southeastern region					−0.0028	−0.0304***
Southeastern region					[−0.56]	[−4.71]
No. of observations	70,038	70,038	70,038	70,038	70,038	70,038
Log-likelihood	−41,851.51	−41,852.17	−38,936.2	−38,933.8	−38,853.7	−38,854.1
IV	No	Yes	No	Yes	No	Yes
Model χ²	10.64	8.68	5071.94	5096.03	5211.60	5196.13
p value	0.0011	0.0032	0.0000	0.0000	0.0000	0.0000
Area under the ROC curve	0.5015	0.5895	0.6777	0.6783	0.6806	0.6834
Wald exogeneity χ²		10.02		52.42		14.77
p value		0.0015		0.0000		0.0001
Convergence	Yes	Yes	Yes	Yes	Yes	Yes

Source: By authors based on ENIGH (National Household Income and Expenditures Survey) 2016.

Note: ROC = receiver operating characteristic. The asterisks denote statistical significance of the coefficient estimates at 1% (***), 5% (**), and 10% (*); z statistics are given in brackets. All estimates were performed using robust standard errors. The coefficients refer to marginal effects. The base category is the traditional region of migration.

Models in columns (3) and (4) control for the households’ demographic and socioeconomic characteristics. Looking at the demographic parameters, we see that the variables related to the presence of children in the household have an important effect on the probability of household spending on domestic tourism; both variables are statistically significant. These results are consistent with those found by Rashidi and Koo (2016), Alegre, Mateo, and Pou (2009, 2013), and Alegre and Pou (2004), who point out that an increase in household size is linked to a reduction in tourism spending. On the other hand, among the demographic characteristics, gender is not found to have a significant influence on tourism expenditure, which is in line with the result presented by Marrocu, Paci, and Zara (2015) and Y. Wang et al. (2006). As for age, it was found that as age increases, the probability of household spending on domestic tourism decreases. This contradicts the findings of previous studies (Alegre, Mateo, and Pou 2009, 2013; Alegre and Pou 2004). Nevertheless, it is worth nothing that the variable squared age of household head is positive, indicating that tourism expenses increase as the head of household ages.

In terms of socioeconomic characteristics, we found that an increase in the head of household’s years of schooling is strongly and positively linked to the probability of household spending on domestic tourism. In fact, each additional year of education completed by the household head increases this probability by approximately 2%. These findings are consistent with studies that have found significant results for education-related variables (Brida and Scuderi 2013; Alegre, Mateo, and Pou 2013). Looking at household poverty, we found that the probability of a household spending on domestic tourism decreases by approximately 5% if the household receives a monetary transfer from a public program. A negative impact on this probability occurs as the number of economic dependents increases. In contrast, the probability of a household’s spending on domestic tourism increases as the number of income earners increases. An additional income earner means an increase of about 3% in the probability of spending on domestic tourism; this result contradicts the findings by Alegre, Mateo, and Pou (2009) and Alegre and Pou (2004). As for home ownership, we found that owning a house increases the probability of household spending on domestic tourism by 1%, consistent with the work of Alegre, Mateo, and Pou (2009, 2013). In addition, homes where one member owns a cell phone are more likely to spend on domestic tourism.

Regarding the effect of international remittances, it continues to be positive and statistically significant. Yet, once again the model in column (3) does not consider the endogeneity of remittances. When endogeneity is considered, column (4), the effect of remittances on the probability of household spending on domestic tourism, increases considerably. In fact, the coefficient indicates that receiving 100 pesos in international remittances is related to a 1.3% increase in the probability of household spending on domestic tourism. The remaining control variables maintain the same sign and level of statistical significance.

Finally, columns (5) and (6) add four regional variables to the model. The purpose of including these variables, as mentioned above, is to capture the heterogeneity of remittance reception across Mexico. In this regard, our results show that living in an urban area compared with a rural area increases the probability of household spending on domestic tourism by 3.7%, statistically different from zero at 1%. Similarly, households located in a region other than the traditional region are less likely to spend on tourism. The coefficients in column (6) show again that international remittances are a fundamental driver of DTE.

It should be highlighted that our estimations for the IV models (columns 2, 4, and 6) are consistent across the three specifications. In these models, we have incorporated the most relevant variables that previous studies have used as tourism expenditure determinants. However, in light of our results, these works have failed to include one factor that can positively influence the probability of household spending on domestic tourism: international remittances.

IV Tobit Results

Once we analyzed the impact of international remittances on the probability of household spending on domestic tourism, we focused on studying the effect that the amount of international remittances received by households has on the amount of their tourism expenditures. To this end, we use a Tobit model (Table 5), and as in the previous analysis, we start with a model with no controls and no correction for endogeneity (column 1). The coefficient indicates that the amount of international remittances is positively associated with tourist expenses. Column (2) shows the same model but uses the instrumental variable (sum_weight_dist); that is, it controls for the endogeneity problem, although the coefficient is not statistically different from zero.

Table 5.

International Remittances’ Impact on DTE.

Dependent Variable: Domestic Tourism Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
International remittances	0.0098**	0.0285	0.0152***	0.2408***	0.0151***	0.3499***
International remittances	[2.53]	[0.40]	[4.46]	[3.76]	[4.37]	[6.98]
Gender of household head			18.512	6.574	27.013**	13.990
Gender of household head			[1.39]	[0.50]	[2.03]	[1.07]
Age of household head			−9.171***	−9.148***	−9.519***	−9.329***
Age of household head			[−4.29]	[−4.27]	[−4.45]	[−4.35]
Squared age of household head			0.1490***	0.1497***	0.1463***	0.1451***
Squared age of household head			[7.09]	[7.11]	[6.97]	[6.90]
Completed years of schooling			87.631***	87.562***	84.66***	84.723***
Completed years of schooling			[33.47]	[33.44]	[32.75]	[32.73]
One-person household			13.721	10.165	14.497	9.1306
One-person household			[0.65]	[0.48]	[0.69]	[0.43]
One child in the household			97.799***	95.619***	93.954***	91.173***
One child in the household			[4.46]	[4.36]	[4.29]	[4.16]
Two or more children in the household			−26.853	−28.507	−22.663	−23.980
Two or more children in the household			[−0.93]	[−0.99]	[−0.79]	[−0.83]
Number of income earners			106.013***	105.762***	102.651***	102.295***
Number of income earners			[20.15]	[20.10]	[19.56]	[19.52]
Number of economic dependents			−66.205***	−64.782***	−67.842**	−66.609**
Number of economic dependents			[−5.96]	[−5.82]	[−6.11]	[−5.99]
Poor household			−209.212***	−210.691***	−180.696***	−176.397***
Poor household			[−14.57]	[−14.60]	[−12.49]	[−12.22]
Housing ownership			39.388***	39.085***	54.431***	57.292***
Housing ownership			[3.07]	[3.05]	[4.23]	[4.44]
Cell phone in the household			281.176***	281.913***	268.414***	262.434***
Cell phone in the household			[16.54]	[16.54]	[15.71]	[15.34]
Area of residence					119.788***	118.842***
Area of residence					[9.33]	[9.27]
Northern region					−67.896***	−76.217***
Northern region					[−4.58]	[−5.18]
Central region					−105.644***	−118.301***
Central region					[−6.18]	[−6.98]
Southeastern region					−67.181***	−172.287***
Southeastern region					[−4.39]	[−8.56]
No. of observations	70,038	70,038	70,038	70,038	70,038	70,038
Log-pseudolikelihood	−221,122.8	−221,137.4	−217,551.8	−217,588.8	−217,489.8	−217,508.2
IV	No	Yes	No	Yes	No	Yes
$F$ test	6.39	0.16	90.00	89.89	69.54	69.44
p value	0.0115	0.6902	0.0000	0.0000	0.0000	0.0000
Wald exogeneity χ²		0.06		15.13		49.73
p value		0.8061		0.0001		0.0000
Convergence	Yes	Yes	Yes	Yes	Yes	Yes

Source: By author based on ENIGH (National Household Income and Expenditures Survey) 2016.

Note: The asterisks denote statistical significance of the coefficient estimates at 1% (***), 5% (**), and 10% (*); z statistics are given in brackets. All estimates were performed using robust standard errors. The coefficients refer to marginal effects. The base category is the traditional region of migration.

In columns (3) and (4), we control for the same group of household demographic and socioeconomic characteristics. Overall, these findings support those found using the Probit model. Among the demographic characteristics, gender is not found to have a significant influence on DTE. However, the sign of the coefficient coincides with that from Bilgic et al.’s (2008) research. On the other hand, as with the Probit model, the coefficient for the age variable is negative and significantly different from zero at 1%. In this case, the coefficient’s sign contradicts the results presented by Bilgic et al. (2008) for tourist spending. Again, if we analyze this variable’s behavior together with that of the squared age of household head variable, we can conclude that as the head of household becomes older the family spends more on domestic tourism. Only one variable related to household size maintains the same algebraic sign and level of statistical significance (one child in the household). This result is consistent with Barquet et al. (2011) that the presence of children younger than 12 years increases tourist expenses, but contradicts the results presented by Zheng and Zhang (2013).

Regarding the socioeconomic variables, the results shown in Table 5 are in line with those found with the Probit model. For example, if we focus on education, the findings in column (4) indicate that for each additional year of schooling the household head completes, DTE increases by 87 pesos (statistically significant at 1%). This result highlights that education has an important effect not only on the probability of household spending on tourism but also on the amount of expenditure incurred. Our findings are consistent with those presented by Zheng and Zhang (2013) and Bilgic et al. (2008) but contradict those of Kim, Han, and Chon (2008), who find a negative relationship between a higher level of education and higher tourism spending. Moreover, the five variables related to the household’s liquidity, as well as to their accumulated wealth (number of income earners, number of economic dependents, poor household, home ownership, and cell phone), have the same algebraic sign and level of statistical significance as the one reported in Table 4.

An additional income earner means an increase of about 106 pesos in domestic tourism expenses; while one more economic dependent decreases household DTE by approximately 65 pesos. Additionally, if a household receives any kind of monetary transfer from a public program, DTE decreases by approximately 210 pesos. In contrast, if any member of the household has a cell phone, spending on domestic tourism increases by about 280 pesos. The model in column (4) considers the endogeneity of remittances, and the coefficient for international remittances indicates that, on average a one-peso increase in the amount of international remittances received raises domestic tourism spending by 0.24 pesos (statistically different from zero at 1%). Compared to the findings in column (3), when considering the endogeneity of remittances, there is a considerably stronger effect.

Finally, columns (5) and (6) add regional controls. The results show that living in an urban area, compared to a rural one, increases household spending on domestic tourism. This result is consistent with that found by Bilgic et al. (2008). Similarly, a household’s domestic tourism spending decreases if it is located in a region other than the traditional region. Regarding the impact of our main variable (international remittances), this remains positive and statistically significant at 1%.

Overall, the estimations in which an instrumental variable approach is used (columns 2, 4 and 6) point toward a high statistically significant effect of international remittances on DTE, which is in line with studies indicating that remittances have a positive impact on the development of migrant-sending countries (Saadi 2020; J. E. Taylor and Castelhano 2016; Kumar 2014; Woodruff and Zenteno 2007; Acosta et al. 2008), thereby improving the living standards and welfare of left-behind families (Berloffa and Giunti 2019; Miftah and Bouoiyour 2015; De Haas 2006; Adams and Page 2005; Koc and Onan 2004), and even positively affecting their subjective well-being, as recent research has shown (Ivlevs, Nikolova, and Graham 2019).⁸

Concluding Remarks

In Mexico, tourism has become one of the most important sectors in terms of both production and income, as well as a source of employment generation and foreign currency. An essential factor in the growth of this sector is the increase recorded in domestic tourism spending, which can be attributed to different factors, among which international remittances stand out as an additional source of income. International remittances comprise a vital source of income for receiving households, as they provide liquidity to cover a variety of expenses ranging from educational services to leisure activities (tourism). In this study we present evidence that international remittances represent an important source of income for Mexican households, particularly in terms of boosting their domestic tourism spending.

Using an instrumental variable Probit and Tobit regression, we analyzed the impact of international remittances on DTE using a sample of Mexican households taken in 2016. Controlling for the endogeneity of remittances, we found that international remittances have a positive and statistically significant effect on the probability of household spending on domestic tourism, meaning that international remittances are an important determinant of tourism spending. Likewise, the results of the Tobit model show that the impact of international remittances on the amount that Mexican households spend on tourism is positive and statistically significant. Thus, compared to benchmark models, the proposed instrumental variable econometric models—which consider the endogeneity of remittances—significantly improve the estimates of international remittances’ causal effect on DTE in terms of both statistical accuracy and robustness of analysis.

Although we recognize that the instrumental variable approach may have some limitations, our findings strongly suggest that international remittances are key to boosting domestic tourism expenditure. Information provided in this study can also help to develop public policies that encourage migrants to spend international remittances on tourism activities, especially in the context of the global economic and health crisis the COVID-19 pandemic has provoked. However, this does not mean that remittances should be used exclusively on tourism-related activities, since it is clear that for some households they are even more important in dealing with other financial constraints (e.g., education, health, and housing). The results of this research highlight that international remittances are a key element of social tourism programs in developing countries. To ensure an increase in the benefits of international remittances, public policies that link tourism activities and remittances to economic growth must be created.

Although this research provides novel findings, two main limitations have been identified (which can guide new research efforts): (1) The impact of international remittances on DTE was computed for only a cross-sectional sampling and (2) the data contained in the survey does not allow a random allocation of international remittances among households. For the former, we recommend that future research use panel data analysis in order to strengthen the findings of this article. In regard to the second point (no random allocation), we suggest the creation of databases that have the main objective of examining the behavior of the migratory phenomenon, and therefore that of the reception of international remittances. Neither limitation reduces the value of our empirical findings, but they do allow us to point out that our findings are not exhaustive, and that alternative methodological approaches and new data sets have the potential to obtain improved results.

Supplemental Material

online_supplementary_material_International_Remittances_as_a_Driver_of_DTE – Supplemental material for International Remittances as a Driver of Domestic Tourism Expenditure: Evidence from Mexico

Supplemental material, online_supplementary_material_International_Remittances_as_a_Driver_of_DTE for International Remittances as a Driver of Domestic Tourism Expenditure: Evidence from Mexico by Jorge Mora-Rivera and Fernando García-Mora-Mora in Journal of Travel Research

Footnotes

Declaration of Conflicting Interests

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

Funding

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

ORCID iD

Jorge Mora-Rivera

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Jorge Mora-Rivera is a Profesor at Tecnologico de Monterrey in Mexico City and a member of the Mexican National System of Researchers. His research mainly focuses on international migration, remittances impacts on development, and tourism economics, with special emphasis on the remittances-tourism nexus.

Fernando García-Mora is a Public Official in the National Council for the Evaluation of Social Development Policy (CONEVAL) in Mexico. Mathematician by the National Polytechnic Institute and MSc in economics from El Colegio de Mexico with a focus on econometrics. He is also a Lecturer at Tecnologico de Monterrey. Most of his research centers on development economics, especially in poverty and its relation to tourism, microfinance, and ICTs.

References

Acosta

Calderón

Fajnzylber

Lopez

2008. “What Is the Impact of International Remittances on Poverty and Inequality in Latin America?” World Development 36 (1): 89–114.

Adams

2011. “Evaluating the Economic Impact of International Remittances on Developing Countries Using Household Surveys: A Literature Review.” Journal of Development Studies 47 (6): 809–28.

Adams

Page

2005. “Do International Migration and Remittances Reduce Poverty in Developing Countries?” World Development 33 (10): 1645–69.

Alegre

Pou

Ll.

2004. “Micro-economic Determinants of the Probability of Tourism Consumption.” Tourism Economics 10 (2): 125–44.

Alegre

Mateo

Pou

Ll.

2009. “Participation in Tourism Consumption and the Intensity of Participation: An Analysis of Their Socio-demographic and Economic Determinants.” Tourism Economics 15 (3): 531–46.

Alegre

Mateo

Pou

Ll.

2013. “Tourism Participation and Expenditure by Spanish Households: The Effects of the Economic Crisis and Unemployment.” Tourism Management 39:37–49.

Ambler

Aycinena

Yang

2015. “Channeling Remittances to Education: A Field Experiment among Migrants from El Salvador.” American Economic Journal 7 (2): 207–32.

Amemiya

1984. “Tobit Models: A Survey.” Journal of Econometrics 24 (1/2): 3–61.

Angrist

Krueger

2001. “Instrumental Variables and the Search for Identification: From Supply and Demand to Natural Experiments.” Journal of Economic Perspectives 15 (4): 69–85.

10.

Azizi

2018. “The Impacts of Workers’ Remittances on Human Capital and Labor Supply in Developing Countries.” Economic Modelling 75:377–96.

11.

Balli

H. O.

Louis

R. J.

2016. “The Impacts of Immigrants and Institutions on Bilateral Tourism Flows.” Tourism Management 52:221–29.

12.

Banxico. 2020. “Sistema de Información Económica: Ingresos por Remesas.” https://www.banxico.org.mx/SieInternet/consultarDirectorioInternetAction.do?sector=1&accion=consultarCuadro&idCuadro=CE81&locale=es (accessed May 20, 2020).

13.

Barquet

Brida

J. G.

Osti

Schubert

2011. “An Analysis of Tourists’ Expenditure on Winter Sports Events through the Tobit Censorate Model.” Tourism Economics 17 (6): 1197–217.

14.

BBVA Research. 2020. “Remittances to Mexico Could Fall 17% in 2020 and Recover between 2023-2028 Due to Covid-19.” https://www.bbvaresearch.com/en/publicaciones/remittances-to-mexico-could-fall-17-in-2020-and-recover-between-2023-2028-due-to-covid-19/ (accessed May 28, 2020).

15.

Berloffa

Giunti

2019. “Remittances and Healthcare Expenditure: Human Capital Investment or Responses to Shocks? Evidence from Peru.” Review of Development Economics 23 (4): 1540–61.

16.

Bettin

Lucchetti

Zazzaro

2012. “Endogeneity and Sample Selection in a Model for Remittances.” Journal of Development Economics 99 (2): 370–84.

17.

Bilgic

Florkowski

W. J.

Yoder

Schreiner

D. F.

2008. “Estimating Fishing and Hunting Leisure Spending Shares in the United States.” Tourism Management 29 (4): 771–82.

18.

Boto-García

Baños-Pino

J. F.

Álvarez

2019. “Determinants of Tourists’ Length of Stay: A Hurdle Count Data Approach.” Journal of Travel Research 58 (6): 977–94.

19.

Brida

J. G.

Cortes-Jimenez

Pulina

2016. “Has the Tourism-Led Growth Hypothesis Been Validated? A Literature Review.” Current Issues in Tourism 19 (5): 394–430.

20.

Brida

J. G.

Lanzilotta

Moreno

Santiñaque

2018. “A Non-linear Approximation to the Distribution of Total Expenditure Distribution of Cruise Tourists in Uruguay.” Tourism Management 69:62–80.

21.

Brida

J. G.

Scuderi

2013. “Determinants of Tourist Expenditure: A Review of Microeconometric Models.” Tourism Management Perspectives 6:28–40.

22.

Buckley

Shekari

Mohammadi

Azizi

Ziaee

2020. “World Heritage Tourism Triggers Urban–Rural Reverse Migration and Social Change.” Journal of Travel Research 59 (3): 559–72.

23.

Cameron

A. C.

Trivedi

P. K.

2005. Microeconometrics: Methods and Applications. New York: Cambridge University Press.

24.

Cameron

A. C.

Trivedi

P. K.

2010. Microeconometrics using Stata: Revised Edition, 2nd ed. College Station, TX: Stata Press.

25.

Cerón

Mora

2014. “El impacto de las remesas internacionales sobre el consumo turístico interno de México, un análisis econométrico con encuestas a hogares.” Turismo y Sociedad 15:75–99.

26.

CICOTUR (Centro de Investigación y Competitividad Turística Anáhuac). 2020. Estimación de las afectaciones al turismo mexicano en 2020 como consecuencia de la pandemia de COVID-19. Documento 14. https://www.anahuac.mx/mexico/cicotur/sites/default/files/2020-05/Doc14_Cicotur_Estimacion_afectaciones_turismo__mexicano_Covid19.pdf (accessed May 15, 2020).

27.

Cohen

J. H.

Rodriguez

2005. “Remittance Outcomes in Rural Oaxaca, Mexico: Challenges, Options and Opportunities for Migrant Households.” Population, Space and Place 11 (1): 49–63.

28.

CONAPO (Consejo Nacional de Población). 2010. “Regiones de origen y destino de la migración México-Estados Unidos.” http://www.conapo.gob.mx/swb/CONAPO/Indices_de_intensidad_migratoria_Mexico-Estados_Unidos_2010 (accessed October 15, 2019).

29.

Cox

Jimenez

1992. “Social Security and Private Transfers in Developing Countries: The Case of Peru.” World Bank Economic Review 6 (1): 155–69.

30.

Demirgüç-Kunt

Córdova

Pería

M. S.

Woodruff

2011. “Remittances and Banking Sector Breadth and Depth: Evidence from Mexico.” Journal of Development Economics 95 (2): 229–41.

31.

De Haas

. 2006. “Migration, Remittances and Regional Development in Southern Morocco.” Geoforum 37 (4): 565–80.

32.

Durand

2016. Historia mínima de la migración México-Estados Unidos. México: Colegio de México. https://www-jstor-org-443.web.bisu.edu.cn/stable/j.ctt1t89k3g.

33.

Dwyer

Seetaram

Forsyth

King

2014. “Is the Migration-Tourism Relationship Only about VFR?” Annals of Tourism Research 46 (3): 130–43.

34.

Etzo

2016. “The Impact of Migration on Tourism Demand: Evidence from Japan.” MPRA, Working Paper No. 72457.

35.

Etzo

Massidda

Piras

2014. “Migration and Outbound Tourism: Evidence from Italy.” Annals of Tourism Research 48:235–49.

36.

Faber

Gaubert

2019. “Tourism and Economic Development: Evidence from Mexico’s Coastline.” American Economic Review 109 (6): 2245–93.

37.

Feng

Page

S. J.

2010. “An Exploratory Study of the Tourism, Migration–Immigration Nexus: Travel Experiences of Chinese Residents in New Zealand.” Tourism Geographies 3 (3): 246–81.

38.

Gao

Wang

2019. “Does High-Speed Rail Boost Tourism Growth? New Evidence from China.” Tourism Management 72:220–31.

39.

Gheasi

Nijkamp

Rietveld

2011. “Migration and Tourist Flows.” In Tourism Economics: Impact Analysis, edited by Matias

Á.

Nijkamp

Sarmento

, 111–26. Berlin: Springer Verlag.

40.

Gullette

G. S.

2009. “Transnational Participatory Development: Economic and Cultural Flows in Oaxaca, Mexico.” Urban Anthropology & Studies of Cultural Systems & World Economic Development 38 (2/4): 235–67.

41.

Gunter

2018. “What Makes and Airbnb Host a Superhost? Empirical Evidence from San Francisco and the Bay Area.” Tourism Management 66:26–37.

42.

Haug

Graham

M. S.

Mehmetoglu

2007. “Little Norway in Spain: From Tourism to Migration.” Annals of Tourism Research 34 (1): 202–22.

43.

Heckman

1997. “Instrumental Variables: A Study of Implicit Behavioral Assumptions Used in Making Program Evaluations.” Journal of Human Resources 32 (3): 441–62.

44.

INEGI (Instituto Nacional de Estadística y Geografía). 2013. “Sistema de Cuentas Nacionales de México. Fuentes y Metodologías. Cuenta Satélite del Turismo de México.” https://www.inegi.org.mx/contenidos/programas/turismo/2013/metodologias/SCNM_Metodo_CSTM_B2013.pdf. (accessed October 14, 2019).

45.

INEGI (Instituto Nacional de Estadística y Geografía). 2017a. “Encuesta Nacional de Ingresos y Gastos de los Hogares 2016. ENIGH. Nueva Serie. Diseño conceptual y definición de categorías y variables.” https://www.inegi.org.mx/app/biblioteca/ficha.html?upc=702825091934. (accessed October 14, 2019).

46.

INEGI (Instituto Nacional de Estadística y Geografía). 2017b. “Encuesta Nacional de Ingresos y Gastos de los Hogares 2016. ENIGH. Nueva Serie. Diseño Muestral.” https://www.inegi.org.mx/app/biblioteca/ficha.html?upc=702825091972. (accessed October 12, 2019).

47.

INEGI (Instituto Nacional de Estadística y Geografía). 2019. “Cuenta Satélite del Turismo de México, 2018.” https://www.datatur.sectur.gob.mx/SitePages/ProductoDestacado3.aspx. (accessed October 20, 2019).

48.

Ivlevs

Nikolova

Graham

2019. “Emigration, Remittances, and the Subjective Well-Being of Those Staying Behind.” Journal of Population Economics 32 (1): 113–51.

49.

Jalil

Mahmood

Idrees

2013. “Tourism-Growth Nexus in Pakistan: Evidence from ARDL Bound Tests.” Economic Modelling 35 (C): 185–91.

50.

Johnston

DiNardo

1997. Econometric Methods. New York: McGraw-Hill.

51.

Koo

T. T. R.

Lim

Dobruszkes

2017. “Causality in Direct Air Services and Tourism Demand.” Annals of Tourism Research 67:67–77.

52.

Kim

S. S.

Han

Chon

2008. “Estimation of the Determinants of Expenditures by Festival Visitors.” Tourism Analysis 13 (4): 387–400.

53.

Koc

Onan

2004. “International Migrants’ Remittances and Welfare Status of the Left-Behind Families in Turkey.” International Migration Review 38 (1): 78–112.

54.

Kumar

R. R.

Naidu

Kumar

2011. “Exploring the Nexus between Trade, Visitor Arrivals, Remittances and Income in the Pacific: A Study of Vanuatu.” Acta Universitatis Danubius, OEconomica 7 (4): 199–218.

55.

Kumar

R. R.

Stauvermann

P. J.

Patel

Prasad

2018. “The Effect of Remittances on Economic Growth in Kyrgyzstan and Macedonia: Accounting for Financial Development.” International Migration 56 (1): 95–126.

56.

Kumar

R. R.

2014. “Exploring the Nexus between Tourism, Remittances and Growth in Kenya.” Quality & Quantity 48 (3): 1573–88.

57.

Lim

Zhu

Koo

T. T. R.

2018. “Urban Redevelopment and Tourism Growth: Relationship between Tourism Infrastructure and International Visitor Flows.” International Journal of Tourism Research 21 (2): 187–96.

58.

Lim

Basnet

H. C.

2017. “International Migration, Workers’ Remittances and Permanent Income Hypothesis.” World Development 96:438–50.

59.

Lima

Eusébio

Varum

C. A.

2018. “Determinants of Leisure Expenditure: Do Low Income Families Behave Any Differently from Others?” Anatolia 29 (1): 24–39.

60.

Lopez-Cordova

2006. “Globalization, Migration and Development: The Role of Mexican Migrant Remittances.” Inter-American Development Bank, Working Paper No. 20.

61.

López-Feldman

Chávez

2017. “Remittances and Natural Resource Extraction: Evidence from Mexico.” Ecological Economics 132:69–79.

62.

Lucas

R. E. B.

Stark

1985. “Motivations to Remit: Evidence from Botswana.” Journal of Political Economy 93 (5): 901–18.

63.

Marrocu

Paci

Zara

2015. “Micro-economic Determinants of Tourist Expenditure: A Quantile Regression Approach.” Tourism Management 50:13–30.

64.

Marschall

2017a. “Migrants on Home Visits: Memory, Identity and a Shifting Sense of Self.” International Journal of Tourism Research 19 (2): 214–22.

65.

Marschall

2017b. “Transnational Migrant Home Visits as Identity Practice: The Case of African Migrants in South Africa.” Annals of Tourism Research 63:140–50.

66.

Massidda

Etzo

Piras

2014. “Migration and Inbound Tourism: An Italian Perspective.” Current Issues in Tourism 18 (12): 1152–71.

67.

McKenzie

Stillman

Gibson

2010. “How Important Is Selection? Experimental vs. Non-experimental Measures of the Income Gains from Migration.” Journal of the European Economic Association 8 (4): 913–45.

68.

McKenzie

Rapoport

2007. “Network Effects and the Dynamics of Migration and Inequality: Theory and Evidence from Mexico.” Journal of Development Economics 84 (1): 1–24.

69.

McKenzie

Yang

2010. “Experimental Approaches in Migration Studies.” World Bank Policy Research Working Paper, Washington, DC: World Bank.

70.

Miftah

Bouoiyour

2015. “The Impact of Migrant Workers’ Remittances on the Living Standards of Families in Morocco: A Propensity Score Matching Approach.” Migration Letters 12 (1): 13–27.

71.

Mitra

Wilson

Managi

Kler

Prayaga

Khanal

2019. “What Determines Whale Watching Tourists’ Expenditure? A Study from Hervey Bay, Australia.” Tourism Economics 25 (7): 1134–41.

72.

Mora

Arellano

2016. “Remittances as Expenditure Drivers in Rural Mexico.” Estudios Fronterizos 17 (33): 231–59.

73.

Mora-Rivera

Cerón-Monroy

García-Mora

2019. “The Impact of Remittances on Domestic Tourism in Mexico.” Annals of Tourism Research 76: 36–52. https://doi.org/10.1016/j.annals.2019.03.002.

74.

Mudarra-Fernández

A. B.

Carrillo-Hidalgo

Pulido-Fernández

J. I.

2019. “Factors Influencing Tourist Expenditure by Tourism Typologies: A Systematic Review.” Anatolia 30 (1): 18–34.

75.

Munshi

2003. “Networks in the Modern Economy: Mexican Migrants in the United States Labor Market.” Quarterly Journal of Economics 118 (2): 549–97.

76.

Ngo

Tsui

K. W. H.

2020. “A Data-Driven Approach for Estimating Airport Efficiency under Endogeneity: An Application to New Zealand Airports.” Research in Transportation Business & Management. Published Online January 21. https://doi.org/10.1016/j.rtbm.2019.100412.

77.

Nunkoo

Seetanah

Jaffur

Z. R. K.

Moraghen

P. G. W.

Sannassee

R. V.

2020. “Tourism and Economic Growth: A Meta-regression Analysis.” Journal of Travel Research 59 (3): 404–23.

78.

Ohe

2020. “Factors on the Linkage between Local Brand Farm Products and Tourism in an Agrarian Community.” In Community-Based Rural Tourism and Entrepreneurship, edited by Ohe

, 217–32. Singapore: Springer.

79.

Ohe

Kurihara

2013. “Evaluating the Complementary Relationship between Local Brand Farm Products and Rural Tourism: Evidence from Japan.” Tourism Management 35:278–83.

80.

Paramati

S. R.

Alam

Md. S.

Chen

C.-F.

2016. “The Effects of Tourism on Economic Growth and CO₂ Emissions: A Comparison between Developed and Developing Economies.” Journal of Travel Research 56 (6): 712–24.

81.

Pardo

A. M.

Dávila

C. A.

2017. “Migración y desarrollo. Características de los hogares y uso de las remesas internas e internacionales en México.” Revista Desarrollo y Sociedad 78:113–41.

82.

Parida

Bhardwaj

Chowdhury

J. T.

2017. “Determinants of Tourism in Indian States: An Empirical Analysis.” Tourism Review 72 (3): 330–43.

83.

Parida

Bhardwaj

Chowdhury

J. T.

2018. “Effects of Economic Development and Crime on Tourism in India.” Anatolia 29 (1): 63–73.

84.

Park

Woo

Nicolau

J. L.

2020. “Determinant Factors of Tourism Expenses.” Journal of Travel Research 59 (2): 267–80.

85.

Pulido-Fernández

J. I.

Cárdenas-García

P. J.

2020. “Analyzing the Bidirectional Relationship between Tourism Growth and Economic Development.” Journal of Travel Research. Published Online June 9, 2020.

86.

Rapoport

Docquier

2006. “The Economics of Migrants’ Remittances.” In Handbook of the Economics of Giving, Altruism and Reciprocity, edited by Kolm

S.-C.

Ythier

J. M.

, 1135–98. North Holland: Elsevier.

87.

Rashidi

T. H.

Koo

T. T. R.

2016. “An Analysis on Travel Party Composition and Expenditure: A Discrete-Continuous Model.” Annals of Tourism Research 56:48–64.

88.

Reyes

R. G.

Mata

N. D.

Gijón-Cruz

A. S.

Cruz

A. C.

López

2009. “Impacto del turismo nostálgico y las remesas familiares en el desarrollo de comunidad rural oaxaqueña.” Migración y desarrollo 12:69–88.

89.

Saadi

. 2020. “Remittance Inflows and Export Complexity: New Evidence from Developing and Emerging Countries.” Journal of Development Studies Published online May 6, 2020. https://doi.org/10.1080/00220388.2020.1755653.

90.

Santos

Vieira

J. C.

2012. “An Analysis of Visitors’ Expenditures in a Tourist Destination: OLS, Quantile Regression and Instrumental Variable Estimators.” Tourism Economics 18 (3): 555–76.

91.

Seetaram

2012. “Immigration and International Inbound Tourism: Empirical Evidence from Australia.” Tourism Management 33 (6): 1535–43.

92.

Song

Shi

Chen

J. M.

Nijkamp

2020. “The Influence of Emigration on Tourism Outward Foreign Direct Investment: Evidence from China.” Journal of Travel Research 59 (4): 631–41.

93.

Stark

1995. Altruism and Beyond: An Economic Analysis of Transfers and Exchanges within Families and Groups. Cambridge: Cambridge University Press.

94.

Taylor

1999. “The New Economics of Labour Migration and the Role of Remittances in the Migration Process.” International Migration 37 (1): 63–88.

95.

Taylor

J. E.

Castelhano

2016. “Economic Impacts of Migrant Remittances.” In International Handbook of Migration and Population Distribution, 525–41. Dordrecht: Springer.

96.

Tobin

1958. “Estimation of Relationships for Limited Dependent Variable.” Econometrica 26 (1): 24–36.

97.

UNWTO (United Nations World Tourism Organization). 2009. Tourism and Migration—Exploring the Relationship between Two Global Phenomena. Madrid, Spain: UNWTO.

98.

UNWTO (United Nations World Tourism Organization). 2019. International Tourism Highlights, 2019 Edition. Madrid, Spain: UNWTO.

99.

UNWTO (United Nations World Tourism Organization). 2020. UNWTO World Tourism Barometer, May 2020—Special focus on the Impact of COVID-19. Madrid, Spain: UNWTO.

100.

Vadean

Randazzo

Piracha

2018. “Remittances, Labour Supply and Activity of Household Members Left-Behind.” Journal of Development Studies 55 (2): 278–93.

101.

Villagómez

2007. “Uso y abuso de las remesas en México.” Acta Republicana: Política y Sociedad 6:31–40.

102.

Wamboye

E. F.

Nyaronga

P. J.

Sergi

B. S.

2020. “What Are the Determinant of International Tourism in Tanzania?” World Development Perspectives 17:100175.

103.

Wang

Chio

D. K.

Law

Fang

2018. “Length of Stay: Its Determinants and Outcomes.” Journal of Travel Research 57 (4): 472–82.

104.

Wang

Davidson

2010. “A Review of Micro-analyses of Tourist Expenditure.” Current Issues in Tourism 13 (6): 507–24.

105.

Wang

Rompf

Severt

Peerapatdit

2006. “Examining and Identifying the Determinants of Travel Expenditure Patterns.” International Journal of Tourism Research 8 (5): 333–46.

106.

Woodruff

Zenteno

2007. “Migration Networks and Microenterprises in Mexico.” Journal of Development Economics 82 (2): 509–28.

107.

Wooldridge

J. M.

2010. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.

108.

World Bank. 2017. Migration and Remittances: Recent Developments and Outlook. Migration and Development Brief No. 28. Washington, DC: World Bank Group.

109.

World Bank. 2020. COVID-19 Crisis through a Migration Lens. Migration and Development Brief No. 32. Washington, DC: World Bank Group.

110.

WTTC (World Travel and Tourism Council). 2019. Travel and Tourism: Economic Impact 2019, World. London, UK: WTTC.

111.

Yang

2008. “International Migration, Remittances and Household Investment: Evidence from Philippine Migrants’ Exchange Rate Shocks.” Economic Journal 118 (528): 591–630.

112.

Yang

2011. “Migrant Remittances.” Journal of Economic Perspectives 25 (3): 129–52.

113.

Zheng

Zhang

2013. “Household Expenditures for Leisure Tourism in the USA, 1996 and 2006.” International Journal of Tourism Research 15 (2): 197–208.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB