Author(s)
Introduction
Remittances, or the money sent from international migrants to family members in their home country, are one of the largest sources of financial flows to developing nations. For some developing countries, remittances have become so economically important that they are as large as their foreign direct investment (FDI) flows.[1] Compared to capital flows, remittance flows tend to be more stable and countercyclical.[2] In 2022 alone, remittances to low- and middle-income countries amounted to $630 billion, or 78% of the total cash transfers worldwide.[3] Top recipient countries include India, Mexico and China, and the largest source of remittances is the United States.[4]
Remittances have accelerated worldwide since 2000, from $123 billion to $773 billion in 2021 (an increase of 528%).[5] Remittances to Mexico, in particular, have increased exponentially since 2013. However, calculating the impact of remittances on economic growth and development is complicated, partly because households in developing countries tend not to use the funds to invest or save. Instead, the money is often used to buy essential commodities such as food, clothing and medical care — essentially alleviating poverty and subsidizing day-to-day expenses.[6] Understanding and forecasting the economic impact of remittances in developing countries is therefore difficult. In the short term, remittances affect an economy by stimulating consumption. In the long term, remittances can impact an economy through investment in physical or human capital. Mexico’s 3x1 Program, for example, was aimed at directing collective remittances to fund community projects and generate long-term economic growth; each Mexican peso contributed by migrants was matched by federal, state and local governments.[7]
Due to the historic growth in the number of remittances sent to Mexico in recent years, recipient households may be consuming more and thereby generating short-term economic growth. But, with increasing dependence on these funds to meet basic life needs and little being reinvested into Mexico's economy, it’s possible that the economy could stagnate in a trend of lower economic growth and higher emigration — often referred to as the “remittance trap.” The International Monetary Fund has warned of the consequences of this trap for developing countries like Mexico.[8] Key questions thus emerge: Do remittances contribute to economic growth and development in recipient countries like Mexico, or do they trap such countries in a cycle of diminished growth? Moreover, given Mexico’s federal makeup, do remittances have differentiated effects in states and municipalities?
This paper uses a simple empirical economic model to examine data on remittances sent to Mexican states and municipalities. The purpose is to investigate whether there is a relationship between remittances and economic growth and development and to understand the importance of remittances for households in Mexico. Interestingly, the findings reveal that although there is no statistical relationship between remittances and economic growth and development at the state level, at the municipal level, remittances appear to have a positive effect on both growth and development. In fact, municipalities that receive remittances have higher economic growth and development compared to those that do not receive remittances. However, Mexico’s economic development should not depend on resources being sent from migrant workers in the United States or elsewhere. Instead, robust economic strategies and policies are needed to boost growth from within.
Literature Review
A number of studies have examined the economic impact of remittances and reached various conclusions. For example, the International Monetary Fund concluded that remittances had a small impact on economic growth in recipient countries, while other studies have found a negative effect.[9] Salahuddin and Gow — who focused on some of the world’s largest remittance recipients including Bangladesh, India, Pakistan and the Philippines — found that there is a significant long-term positive relationship between remittances and economic growth, but there is no significant short-term relationship.[10] Similarly, Carvajal and Almonte found that remittances in Mexico had a positive impact on economic development, but only if credit constraints were relaxed.[11] In general, these studies reveal that measuring and understanding the effect of remittances on economic growth is a difficult process that can produce mixed results.
Other studies have concluded that remittances may affect economic growth and development, but it depends on how households spend their resources — whether they invest in human and physical capital, consume more goods and services, contribute to their savings or use the money merely to survive. If households use remittances to finance productive investments and activities, they can improve their income over time, which is important for marginalized or rural areas in developing countries.[12] Indeed, various researchers have found that remittance flows can help stabilize household income and improve living conditions. For example, according to Amuedo-Dorantes, remittances have helped Mexican households with saving constraints to smooth their income.[13] Remittances can also increase household consumption of goods and services, thereby generating greater aggregate demand and promoting economic growth. Additionally, empirical evidence reveals that remittances contribute to economic growth through a positive impact on investment and consumption.[14] However, by spending remittances on unproductive investments, barriers to economic growth can emerge. The evidence varies by country, but in general, most studies find that remittances boost economic growth[15] and reduce poverty and inequality.[16]
Zahid Hussain of the World Bank establishes that the expenditure of remittances can generate a positive multiplier effect on income and output if there is a linkage between the localities that receive remittances and the national economy. However, when these resources are concentrated in only limited areas, the multiplier effect is not transferred to the national economy.[17] Further research reveals that remittances can reduce incentives to work and lower labor force participation by adversely affecting the labor supply decisions of recipient families.[18] Still other studies have found that rising levels of remittances could be harmful to the long-term growth of recipient economies through an appreciation of the real exchange rate, which makes recipient economies less competitive in international trade.[19]
Households receiving remittances may spend the funds on education or health care, thereby promoting economic development. Most research in developing countries has found evidence of a positive effect of remittances on education and health.[20] Indeed, evidence shows that remittances have a positive correlation with human development levels, especially in middle-income countries. Studies also reveal that the general impact of remittances on both relative and absolute changes in human development indices (HDIs) is significant. In fact, remittances have a higher impact than FDI and public expenditure on the economic development of these countries.[21] Another study that used a sample of low- and middle-income countries found that the relationship between remittances and development outcomes, such as education and health, is robust at the aggregate level after controlling for economic and geographic differences and time trends.[22]
Modeling of the impact of remittances on economic growth ranges from cross-sectional data to panel data.[23] Most of the studies that analyze this relationship use a sample of countries. In the case of Mexico, the analysis focuses more on states and panel data, revealing a positive impact of remittances on economic growth.[24] Little research exists on how remittances impact economic development, specifically, but the few existing studies show that remittances have an important effect on the economic development of Mexican states.[25] Research at the municipal level is particularly scarce, but one study found that development indicators for education and health improve as the number of households that receive remittances increases.[26]
This paper adds an important contribution to the existing literature on remittances by examining their impact on the economic growth and development of Mexican states and municipalities. It is crucial to understand the impact of remittances at these levels, because any negative change in international markets (particularly in the U.S.) could impact the ability of migrant workers to send funds home. This, in turn, would negatively affect not only the local economies in receiving countries, but also people’s welfare.
Methodology
Two empirical models — with slight differences — were used to analyze data on remittances to Mexico. One tests if remittances boost economic growth, and the other tests if these resources drive economic development for Mexican states and municipalities. The first empirical model has a dependent variable — the rate of growth of per capita income — and a set of independent variables, one of which is remittances. In this case, the equation is as follows:
YiT is the real gross domestic product (GDP) per capita in economy i at time T; Yi0 is the real GDP per capita in economy i at time 0; α is a constant; T is the number of years; Wi0,T is the error term; and Z represents different control variables. The key independent variable for this model is remittances[27] as a percentage of GDP. The expected sign (positive or negative) is not clear, since it will depend on how the remittances are used by recipients in Mexico’s states and municipalities. If households use remittances to purchase goods to alleviate poverty and meet basic needs, there would be no effect on economic growth. In some cases, remittances could generate incentives to reduce productive efforts, thereby negatively affecting economic growth. On the contrary, if these households use these resources to invest in physical and human capital, there could be a positive effect on both economic growth and development. In this case, remittances can boost economic growth by increasing aggregate demand.[28]
The variable of trade can be added to the above model. Trade liberalization can have positive effects for both short-term and long-term economic growth.[29] A vital factor in trade liberalization and economic growth is foreign direct investment (FDI). Resources from FDI can be used to buy better technologies that increase companies’ capital stock and efficiency and thereby generate growth. Hence, there should be a positive relationship between FDI and the growth rate of per capita income.[30] In addition, investment can reduce disparities between economies and cause the flow of productive factors to generate economic growth.
The variables used to test this model include: the growth rate of state GDP per capita (dependent variable), initial level of state GDP per capita, proportion of FDI with respect to state GDP, remittances with respect to state GDP, public investment with respect to state GDP, initial state population, and marginalization index by state.[31] These six variables are used as independent variables for the state econometric model; other regional variables are added later.
The second empirical model analyzes state economic development. In this case, GDP is replaced with HDI, and the same independent variables are used except for the initial level of state GDP per capita. In this case, remittances may promote development if recipients spend them on education or health care or invest in entrepreneurial activities.
There are different approaches to estimate the parameters of these models. This paper uses Ordinary Least Squares (OLS) with cross-sectional data. A potential problem of endogeneity of remittances could emerge, since these resources could be correlated with the error term. If that is the case, the estimates produced by equation 1 may be biased. This paper assumes that the control variables included in the models can eliminate this issue.
Regarding the regressions of municipalities, these change slightly due to a lack of data. Instead of using municipal GDP, the gross census value added (VACB) by municipality is used as a proxy variable. The variables added to the growth regression include: the growth rate of municipal VACB per capita (dependent variable), the initial level of municipal VACB per capita, remittances with respect to municipal VACB, public investment with respect to municipal VACB, initial municipal population, marginalization index by municipality, regional variables, and dummy variables that establish if a municipality receives remittances and if it has more than 90% of its population living in poverty. As in the case for states, for the economic development model, VACB is replaced with the municipal HDI, and the independent variables previously mentioned are used.
Data
This section explains the nature of the data used and clarifies some key points on the main and lesser-known variables.
The variable GDP (base year 2013) by state is obtained from the National Institute of Statistics and Geography (INEGI).[32] In the case of municipalities, the variable VACB provided by INEGI is used as a proxy for GDP since there is no data on GDP of the municipalities for the period analyzed.[33] INEGI, however, has provided data on public investment by state and municipality since 1989.[34] The HDI variables by state and municipality are obtained from United Nations Development Programme (UNDP) and Global Data Lab.[35] The HDI is a measurement to evaluate the level of individual human development in a state and a municipality. It combines key variables for its measurement: life expectancy at birth, literacy and school enrollment rates, and GDP per capita.[36] The indices for each of these dimensions are aggregated with equal weights in a simple average to estimate the HDI.
On demographic data, Mexico’s National Population Council (CONAPO) has calculated the index of marginalization (IM) by state and municipality since 1990.[37] IM provides data on Mexican states and municipalities according to the impacts suffered by the population as result of the lack of access to education, adequate housing, sufficient income and the distribution of population in small localities.[38] The National Council for the Evaluation of Social Development Policy (CONEVAL) estimates the population living in poverty and extreme poverty by state and municipality.[39] CONEVAL and INEGI also estimate the population by state and municipality.[40]
Banco de Mexico (BANXICO) estimates the number of remittances by state since 2003 and by municipality since 2013.[41] The Mexican Secretariat of Economy (SE) provides information on FDI by state since 1999, but not for municipalities.[42] Finally, INEGI has generated data on public resource expenditures by state and municipal governments. In this case, it provides information for public investment since 1989 for all states and most municipalities (78%).[43] The econometric models studied Mexican states from 2005 to 2019 and municipalities from 2013 to 2018.[44] In the case of municipalities, a smaller timeframe is used because the information on remittances for municipalities began in 2013.
HDI and IM
Mexico has 32 states and 2,469 municipalities, which are equivalent to counties in the U.S.[45] The state with the highest number of municipalities is Oaxaca with 570, and the state with the lowest number is Baja California Sur with only five. Guerrero, Chiapas and Oaxaca are the states with the lowest HDI compared to other states in recent years. Alternatively, the states with the highest human development have been Mexico City and Nuevo Leon.[46] The municipalities with the lowest HDI have been Cochoapa el Grande (Guerrero), San Martín Peras (Oaxaca), Batopilas (Chihuahua), Santos Reyes Yucuná (Oaxaca) and Coicoyán de las Flores (Oaxaca).[47] Oaxaca has five of the municipalities with the lowest HDI, while Mexico City has six with the highest HDI. The IM also shows Guerrero, Chiapas and Oaxaca with very high marginalization rates between 2005 and 2020. These three states have 10.5 % of the national population (13.2 million people). The fact that these states have had persistently high levels of marginalization over the last 15 years tells the story of a “marginalization trap.” If a state is marginalized, it appears doomed to be so for a long time. The states with the lowest marginalization rates were México City, Nuevo León and Coahuila. These three states represent 14.4% of the total population (18.1 million people). In the case of municipalities, this index shows that one third of municipalities have a high or very high marginalization rate, where approximately 9.1% of the national population live (more than 11.4 million people).[48] The municipalities with the worst economic and social conditions in 2013 and 2020 were Batopilas de Manuel Gómez Morín (Chihuahua), Mezquital (Durango) and Del Navar (Nayarit). The municipalities with low and very low levels of marginalization represent 83% of the total population (more than 105.2 million). These municipalities are located in the center and north of the country and include Benito Juárez (Mexico City), San Nicolás de los Garza (Nuevo León), Cuauhtémoc (Mexico City) and Apodaca (Nuevo León).[49]
Remittances
Remittances sent to Mexico have grown since 2013 (Figure 1), reaching a historic high in 2021 at about $51.6 billion — an increase of 27% over 2020 and 131.3% over 2013. Remittance income as a percentage of the GDP in Mexico has also grown considerably since 2013. In 2021, it made up 4.0% of the country’s GDP, a 2.3 percentage point increase over 2013 (Figure 2). These numbers show how important remittances have been to the Mexican economy in recent years.
Figure 1 – Mexico: Remittances Inflow, 2000-2021 (Billions of Dollars)
Remittances have become so important to Mexico that they are as large as, or larger, than FDI flows to the country (Figure 2). Moreover, whereas remittances have grown in recent years as a percentage of GDP, FDI has decreased since 2017. In 2020, remittances became the second largest source of income in the country, only behind auto and auto parts exports. FDI is in third place.
Figure 2 – Mexico: Income from Remittances and FDI, 2010-2021 (percentage of GDP)
Most remittances to Mexico come from the United States. In 2021, they amounted to more than $48.9 billion (94.9% of the total received in Mexico), with most of the remittances coming from California ($16.2 billion) and Texas ($7.7 billion).[50] The Mexican states that receive the majority of these remittances are those that have sent more migrants to the United States in recent years. The main recipients were Jalisco, Michoacán and Guanajuato. In 2021, Jalisco received the most out of 32 states with a total of $5.2 billion. Its remittances increased 26.1% with respect to 2020.[51] Michoacán was second with a total of $4.9 billion and an increase of 22.9% compared to 2020. Of the Mexican municipalities, Tijuana (Baja California), Guadalajara (Jalisco) and Alvaro Obregon (Mexico City) were the three main recipients of remittances with a combined total of $1.9 billion in 2021.[52]
Remittances represent a major part of state and municipal income. The states with the highest level of remittances as a percentage of their GDP were Michoacán (16%), Guerrero (12%), Zacatecas (10.2%) and Oaxaca (10.1%) in 2005. This distribution is almost identical in 2019: Michoacán (17.1%), Zacatecas (14.8%), Guerrero (14.6%) and Oaxaca (14.5%). In all these cases the percentages increased compared to 2005. Given this, it appears that remittances have become a structural pillar of the economies in these states. The municipalities that received the most remittances in 2013 were Tijuana, Baja California ($327 million); Puebla, Puebla ($345 million); and Guadalajara, Jalisco ($306 million). As with the states, this distribution was similar in 2019 with increasing amounts: Tijuana, Baja California ($481 million); Puebla, Puebla ($461 million); and Morelia, Michoacán ($442 million).
Foreign Direct Investment
The states that received the most FDI in 2005 and 2019 were Mexico City, Nuevo Leon and the State of Mexico.[53] These three states received more than 52% of the total FDI in 2005 and more than 42% in 2019. The states that received the least amount of FDI were Durango, Colima and Zacatecas in 2005 (0.4% of the total FDI), and Campeche, Oaxaca and Colima in 2019 (0.7% of the total FDI).
Public Investment
States in Mexico invest, on average, very little of their public resources as a percentage of their GDP. In 2005, this percentage was 0.63%, and in 2019, it was 0.42%. The states with the highest levels of public investment were Hidalgo, Tlaxcala and Chihuahua with an average public investment with respect to their GDP of 1.2% in 2005. In 2019, those states were Tlaxcala, the State of Mexico and Sinaloa with 0.96%. The states with the lowest levels of public investment were Campeche, Tabasco and Mexico City (on average 0.06% of GDP) in 2005. For 2019, the states were Baja California, Nuevo León and Yucatán (on average 0.06% of GDP).[54]
Results
This section provides the results of the econometric models for both states and municipalities.[55] Based on these results, it is possible to assess whether remittances boost Mexico’s economic growth and economic development at the state and municipal levels.
States
Economic Growth and Remittances
The states with the highest economic growth from 2005 to 2019 were Mexico City (2.0%), Aguascalientes (1.8%), San Luis Potosí (1.7%) and Chihuahua (1.7%). Meanwhile, the states with the lowest economic growth were Campeche (-6.4%), Chiapas (-1.2%), Tabasco (-0.9%) and Morelos (-0.3%). In this period, the growth rate of GDP per capita for all 32 states averaged 0.5%. An interesting case is Campeche. It has the highest GDP per capita of any state. Oil activity likely distorts this measure. However, its level of economic activity has deteriorated significantly. In 2019, its GDP per capita was 59% lower than in 2005, representing an annual average drop of approximately 6.4%.
Figure 3 shows the growth rate of per capita income between 2005 and 2019 relative to the remittances (% of GDP) in 2005 for all Mexican states. The relationship appears to be positive. However, using the model, we can test to see if it is actually positive and statistically significant. In this case, there is an outlier — Campeche — but the results in Table 2, with or without the outlier, are basically the same.[56]
Figure 3 – Mexico: Remittances 2005 (% GDP) and Rate Growth of GDP per capita, 2005-2019
To determine the impact of remittances on the economic growth of Mexican states, we must control for other variables. First, only the variable remittances as a percentage of GDP is added to the model (Model A). This model provides a benchmark for the analysis. Then, other control variables are added (Model B).
Model C includes the regional dummy variables. Adding these variables to the model helps verify differences among Mexican regions.[57] In this case, we use Esquivel’s definition of regions (see Table 1).[58]
Table 1 – Definition of Regions
Table 2 shows the results of Models A, B and C. Model A shows no impact of remittances on the growth rate of per capita income. This result does not change if other variables are added to the model since remittances are not statistically significant. Model B establishes that there is no statistical evidence to support the impact of remittances (as a percentage of GDP) on economic growth. States with high levels of initial income have lower economic growth than states with lower levels of initial income for the analyzed period. The index of marginalization is statistically significant and robust. There is a negative relationship between marginalization and economic growth. States with high levels of marginalization have lower per capita income growth rates. This reinforces the earlier discussion of the marginalization trap and shows that it does exist at the state level.[59]
FDI, as a percentage of GDP, is statistically significant at 10%. At the state level, this variable affects economic growth. The higher a state’s FDI, the higher its per capita income growth rate for the period of study. Finally, population and public investment per capita are not statistically significant. Hence, the resources spent by state governments do not translate into economic growth. Adding all these variables increases the “goodness-of-fit” of the model to 0.7546.
Table 2 – Regression results for the States of Mexico, 2005-2019
In this case, all of the regional variables are not statistically significant, but those that were significant in the previous model remained significant. Given the results in Table 2, remittances do not affect economic growth at the state level for the period analyzed.
Economic Development and Remittances
It is assumed that the previous model can be applied to the model for economic development by substituting GDP with HDI. Given the data, the period will be from 2005 to 2018. During this period, the growth rate of HDI for all states averaged 0.4%, a very low rate that reveals the lack of development in education and health in Mexico’s states.
Figure 4 shows the growth rate of HDI between 2005 and 2018 in relation to the variable of remittances (% of GDP) in 2005 for all Mexican states. As can be seen from the figure, there appears to be a positive relationship between these variables as established by the fitted line — but this has to be tested with the empirical model.
Figure 4 – Mexico: Remittances 2005 (% of GDP) and Rate Growth of HDI 2005-2018
To analyze the impact of remittances on the economic development of Mexican states, in the first specification, only the remittances variable as a percentage of GDP is added to the model (Model D). Then, other control variables are included in the previous case (Models E and F) to provide the main results of these econometric models (Table 3).
The first model shows that remittances have a positive impact on the growth rate of development. However, this result is not robust since adding more control variables makes this variable not statistically significant (Model D). Model E establishes that there is no statistical evidence to support the impact of remittances (as a percentage of GDP) on economic development. States with high levels of initial human development have lower economic development than states with lower levels of initial development for the analyzed period. All other control variables are not statistically significant, except for population. However, if we add the regional dummy variables, they become statistically significant, and the population variable loses significance (Model F). Keeping the initial level of HDI constant, the states belonging to all regions but the South tend to develop faster than states in the southern region for the period analyzed. Some other factors are not included in the model (for example, the abundance of human capital in each state). Given these results, we can say that remittances do not affect economic development at the state level for the period 2005 to 2018.
Table 3 – Regression Results for Mexico’s States, 2005-2018
Municipalities
Economic Growth and Remittances
Given that there is no data on municipal GDP from 2013 to 2018, it is necessary to use a different variable. VACB is a component of GDP, and INEGI establishes that, technically, VACB represents GDP before taxes.[60] So, in this paper, it is assumed that VACB is equivalent to GDP and it is used as a proxy variable. These two variables are highly correlated, and they should provide the same results with respect to the relationship of the variables analyzed. In the case of the states, the Pearson’s correlation coefficient estimated between GDP per capita and VACB per capita is 0.9883 for 2005 and 0.9654 for 2019.[61]
In the period analyzed, municipalities had high volatility in terms of the growth rate of VACB per capita. In this case, if we adjust and estimate the average growth rate of GDP per capita, it would be around 0.8%.
To determine the impact of remittances on the economic growth of Mexican municipalities, first, the remittances variable as a percentage of VACB is added as the only independent variable to the model (Model G), as a benchmark for the analysis. Then, other control variables are included to the previous model. Both municipalities that receive remittances and municipalities that do not receive remittances are considered, so a dummy variable is added depending on if the municipality receives or does not receive remittances (Model H). It is important to mention that there is no information on FDI for municipalities, so this variable cannot be included in this analysis. Finally, Model I is based on the previous model but includes the regional dummy variables.
Table 4 shows the results of Models G, H and I. In the first model, remittances have a positive impact on the growth rate of per capita income. This result is not robust, since adding more control variables changes the statistical significance of remittances. Model H includes more independent variables. In this case, municipalities with high levels of initial income have lower economic growth than municipalities with lower levels of initial income for the analyzed period. The index of marginalization is statistically significant and robust. Again, there is a negative relationship between marginalization and economic growth. Municipalities with high levels of marginalization have lower per capita income growth rates. This further reinforces the earlier discussion of the marginalization trap and shows that it does exist at all levels. The population and public investment (as a percentage of VACB) are positive and statistically significant. The greater the population a municipality has, the greater its economic growth. Additionally, the resources spent on public investment by municipal governments translate into economic growth for the period analyzed. In this case, remittances have no impact on economic growth. However, those municipalities that receive remittances have higher economic growth than those that do not receive remittances. Two more dummy variables are added to this model. The first one assigns a value of 1 if the municipality has more than 90% of its population living in poverty and 0 otherwise.[62] The second variable assigns a value of 1 if the municipality has more than 90% of its population living in poverty and receives remittances, and 0 otherwise.
As seen, these two variables are not statistically significant for municipalities. Given the results on remittances, it can be said that remittances do not have an impact on municipal economic growth for the period analyzed. However, municipalities that receive remittances have higher economic growth compared to those that do not receive remittances. Adding the regional dummy variables do not change the results just mentioned, so in Model I the results are the same as in Model H.
Table 4 – Regression results for the Municipalities of Mexico, 2013-2018
Municipalities by State — Economic Growth
Regarding municipalities by state, the same model cannot be used for all Mexican states since some states do not have enough municipalities to run an econometric model. However, the three states with the highest poverty levels in 2015 are examined as well as those with the highest number of municipalities: Oaxaca, Chiapas and Guerrero. The objective is to know if remittances have an impact on the economic growth of these poor municipalities. So, the variables analyzed are those related to remittances, since the other control variables yield nearly the same results as before. The remittances variable as a percentage of VACB is not significant for the three cases.
For Oaxaca and Chiapas, the municipalities that receive remittances grow faster than those that do not. But there is no statistical evidence for the other dummies. In the case of Guerrero, the dummy variable for remittances alone is not statistically significant. Nevertheless, its municipalities with more than 90% of their population living in poverty and receiving remittances have a higher economic growth rate than those municipalities without these characteristics. In addition, its municipalities with more than 90% of their population living in poverty have lower economic growth than the excluded group.[63]
Economic Development and Remittances
As in the analysis of states, it is assumed that the previous model can be applied to the economic development model for municipalities by substituting GDP with HDI and focusing on the period from 2013 to 2018. The growth rate of HDI for all municipalities averaged 0.06% during this period, a very low rate that reveals the lack of human development in all of Mexico’s municipalities.
Three models are shown in Table 5. The first model states that remittances are not statistically significant (Model J). However, this result changes for the other two cases, as will be explained later. Model K establishes that the index of marginalization is statistically significant and robust. There is a negative relationship between marginalization and economic development at the municipal level. As in the other cases, the marginalization trap exists in this scenario. Municipalities with high levels of initial human development have lower economic development than municipalities with lower levels of initial development for the analyzed period. The population and public investment variables are not statistically significant. So, there is no statistical evidence to affirm that the resources allocated to public investment by municipal governments have an impact on their economic development for the period analyzed. The more remittances (as a percentage of VACB) a municipality receives, the less economic development it has. However, the municipalities that receive remittances develop more than those that do not receive remittances. This means that municipalities that depend more on remittances use those resources to meet the basic needs of their inhabitants and not to promote human development. But on average, municipalities that receive remittances develop more than their counterparts. The other dummy variables are also statistically significant. If municipalities have more than 90% of their population living in poverty, they experience less economic development than other municipalities. The next result is crucial; if municipalities have more than 90% of their population living in poverty and receive remittances, they will experience more economic development than other municipalities. In other words, these poor communities are using these resources for economic development since it is the only way to promote education and health. The results for dummy remittances reinforce their importance in the economic development of Mexican municipalities, particularly in the poorest municipalities that receive these resources.
Adding regional dummy variables to the above model produces the same results, but the result of regional disparities between municipalities is present, at least for the Center, Center-North, Gulf and North (Model L). These variables are positive and statistically significant. So, if a municipality is located in the aforementioned regions, it will develop more than the municipalities that are in the excluded region: the South.
Table 5 – Regression results for the Municipalities of Mexico, 2013-2018
Municipalities by State — Economic Development
Regarding municipalities by state, we analyze the impact of remittances in the same states as before: Chiapas, Oaxaca and Guerrero. In Chiapas, the remittance variable (as a percentage of VACB) is not statistically significant, as it is in Guerrero. In Oaxaca, this variable is negative and statistically significant. Meanwhile, in Chiapas and Oaxaca, the municipalities that receive remittances develop more than those that do not. In the case of Guerrero, this variable is not statistically significant, so remittances have no effect on the economic development of its municipalities. In all these states, there is no statistical evidence that remittances have an impact on the economic development of municipalities with more than 90% of their population living in poverty. Hence, it is possible that poor municipalities in Chiapas and Oaxaca are using remittances only as income to purchase goods and services essential for survival, but not for education and health (i.e., not for their economic development). Recalling the previous section, it can be said that remittances in poor municipalities in Guerrero foster economic growth, but not economic development.
Conclusion
The level of remittances received in Mexico has risen to record levels since 2013. Have these resources impacted economic growth and development in Mexico’s states and municipalities? This study provides information on how recipients of remittances use those resources. If remittances reduce growth and also cause families to have an increased dependence on them, a remittance trap could exist. Although this paper analyzes the relationship between economic growth and remittances, it does not examine whether there has been increased emigration from Mexico, so the presence of this trap cannot be fully verified. This paper also does not consider the spatial correlation between these variables — an element that should be analyzed in future research.
Mexico has had an average economic growth rate of around 1.9% per annum for the last decade — i.e., very low economic growth. From 2019 to 2021, the Mexican economy contracted an average of 1.2%, but remittances could help reduce this contraction if they were used for productive activities. If remittances are used only to satisfy basic needs and not to generate productive activities, they are unlikely to produce growth in the long run. Remittances can also foster economic development if they are used for education and health. But if they are not used for these things, they can negatively impact the economic development of Mexico’s states and municipalities.
This research finds that there is no statistical relationship between remittances (as a percentage of GDP) and economic growth in Mexican states for the period of 2005 to 2019. The same result is obtained for remittances and economic development in Mexican states for the period of 2005 to 2018. These results depend on the methodology used. In the case of Mexican states, there is no statistical evidence to say that remittances affect these economies and their development. However, in the case of municipalities, the results provide important information about how remittances are used. The regression results express a positive relationship between remittances (as a percentage of VACB) and economic growth for Mexican municipalities, but this result is not robust. However, the results for the remittance dummy variable are robust and positive. That is, the municipalities that receive remittances have higher economic growth than those that do not receive remittances for the period of 2013 to 2018. These municipalities are putting their resources not only toward meeting their basic needs, but also toward productive activities that favor economic growth.
In the case of economic development, there is a negative relationship between remittances (as a percentage of VACB) and economic development. The more remittances (as a percentage of VACB) a municipality receives, the less economic development it has. The magnitude of its coefficient is quite small, however. For example, an increase of one unit in the remittances variable (as a percentage of VACB) is associated with a reduction of 0.0018 percent change in economic development. The municipalities that receive the most remittances are not putting them toward education and health. However, on average, the municipalities that receive these resources develop more than those that do not receive remittances. If these municipalities receive large amounts of remittances relative to their GDP, a point is reached where these resources are no longer invested in their economic development. Therefore, it can be concluded that, if a municipality receives remittances, it uses this income to boost economic growth and development. In the case of municipalities with more than 90% of their population living in poverty and receiving remittances, they will develop more than other municipalities. These poor municipalities are allocating their resources to economic development.
This paper finds that in the case of municipalities, remittances appear to have a positive effect on economic growth and development. However, this relationship could be problematic since their economic development is tied to resources being sent from workers in other countries and not on the economic strategies and public policies of the Mexican government. It is not certain that these individuals will be able to continue to send this money to Mexico permanently, despite the fact that this has been the case in recent years. The best way to grow and develop in the long term is to invest in human and physical capital. Therefore, better public policies must be established to improve education, financial inclusion, health and security in Mexico. Mexico also needs a better economic strategy focused on creating new firms and jobs to reduce poverty and inequality. This does not mean that resources from abroad shouldn’t be used, but it is important that they are not relied upon as a key driver of Mexico’s economy.
Endnotes
[1] Adolfo Barajas et al., “Do Workers’ Remittances Promote Economic Growth?” International Monetary Fund, IMF Working Paper, 2009, https://www.imf.org/external/pubs/ft/wp/2009/wp09153.pdf.
[2] Dilip Ratha, “What are Remittances?” International Monetary Fund, Economics Concepts Explained, n.d., https://www.imf.org/external/Pubs/FT/fandd/basics/pdf/ratha-remittances.pdf.
[3] “A War in a Pandemic: Implications of the Ukraine Crisis and COVID-19 on Global Governance of
Migration and Remittance Flows,” World Bank, Knomad, Migration and Development Brief 36, May 12, 2022, https://www.knomad.org/sites/default/files/2022-07/migration_and_development_brief_36_may_2022_0.pdf.
[4] “Mexico: Yearbook of migration and remittances 2022,” BBVA Research, BBVA Foundation and the National Population Council, September 5, 2022, https://www.bbvaresearch.com/en/publicaciones/mexico-yearbook-of-migration-and-remittances-2022/.
[5] “Personal Remittances Received in 1997,” International Organization for Migration's Global Migration Data Analysis Centre (IOM’s GMDAC), Migration Data Portal, https://www.migrationdataportal.org/international-data?t=1997&i=remit_re_excel.
[6] If remittances are used as permanent income rather than temporary income, households tend to consider them more reliable to support daily expenses. See Donghui Wang, Annelise Hagedorn, and Guangqing Chi, “Remittances and Household Spending Strategies: Evidence from the Life in Kyrgyzstan Study, 2011–2013,” Journal of Ethnic and Migration Studies 47, no. 13 (2021): 3015–3036, https://doi.org/10.1080/1369183x.2019.1683442.
[7] Francisco Javier Aparicio and Covadonga Meseguer, “Collective Remittances and the State: The 3x1 Program in Mexican Municipalities,” CIDE, Mexico, July 2009, http://investigadores.cide.edu/aparicio/Aparicio%26Meseguer_CollectiveRemittances3x1_09.pdf; “Mexican migrant program will be strengthened with IDB support,” Inter-American Development Bank, September 20, 2012, https://www.iadb.org/es/noticias/comunicados-de-prensa/2012-09-20/programa-de-migrantes-mexicanos-p3x1%2C10117.html.
[8] Ralph Chami, Ekkehard Ernst, Connel Fullenkamp and Anne Oeking, “Is there a Remittance Trap?” IMF Finance and Development Magazine, September 2018, https://www.imf.org/Publications/fandd/issues/2018/09/is-there-a-remittance-trap-chami.
[9] Barajas et al., “Do Workers’ Remittances Promote Economic Growth?”
[10] Salahuddin and Gow, “The Relationship Between Economic Growth and Remittances.”
[11] Gutierrez and Almonte, “Remesas y crecimiento: un análisis estructural para México.”
[12] Oded Stark, J. Edward Taylor, and Shlomo Yizhaki, “Migration, Remittances and Inequality,” Journal of Development Economics 28, no. 3 (May 1988): 309-322, https://doi.org/10.1016/0304-3878(88)90002-8; Inayah Hidayati, “Migration and rural development: The impact of remittance,” IOP Conference Series: Earth and Environmental Science 561, no. 1 (September 2020): 1-6, https://doi.org/10.1088/1755-1315/561/1/012018.
[13] Catalina Amuedo-Dorantes, “The good and the bad in remittance flows,” IZA World of Labor, November 2014, https://wol.iza.org/uploads/articles/97/pdfs/good-and-bad-in-remittance-flows.pdf?v=1.
[14] Jemma Dridi, Tunc Gursoy, Hector Perez-Saiz, and Mounir Bari, “The Impact of Remittances on Economic Activity: The Importance of Sectoral Linkages,” International Monetary Fund, Working Paper, (August 2019): 5-37, https://www.imf.org/~/media/Files/Publications/WP/2019/wpiea2019175-print-pdf.ashx.
[15] Alina Cazachevici, Tomas Havranek, and Roman Horvath, “Remittances and economic growth: A meta-analysis,” World Development 134 (October 2020), https://doi.org/10.1016/j.worlddev.2020.105021.
[16] Amuedo-Dorantes, “The good and the bad in remittance flows.”
[17] Zahid Hussain, “Can International Remittances Be Unproductive in Recipient Countries? Not Really!” End Poverty in South Asia (World Bank blog), February 16, 2014, https://blogs.worldbank.org/endpovertyinsouthasia/can-international-remittances-be-unproductive-recipient-countries-not-really.
[18] Soma Rani Sutradhar, “The Impact of Remittances on Economic Growth in Bangladesh, India, Pakistan and Sri Lanka,” International Journal of Economic Policy Studies 14, no. 1 (January 31, 2020): 275-295, https://doi.org/10.1007/s42495-020-00034-1.
[19] Dridi, Gursoy, Perez-Saiz, and Bari, “The Impact of Remittances on Economic Activity.”
[20] Maria Cristina Zhunio, Sharmila Vishwasrao, and Eric P. Chiang, “The Influence of Remittances on Education and Health Outcomes: A Cross Country Study,” Applied Economics 44, no. 35 (December 2012): 4605-4616, https://doi.org/10.1080/00036846.2011.593499.
[21] Aysen Ustubici and Darja Irdam, “The impact of remittances on human development: A quantitative analysis and policy implications,” Economics & Sociology 5, no. 1 (May 20, 2012): 74-95, https://doi.org/10.14254/2071-789X.2012/5-1/6.
[22] Zhunio, Vishwasrao, and Chiang, “The Influence of Remittances on Education and Health Outcomes.”
[23] Miguel Ángel Mendoza González and Marcos Valdivia López, “Remesas, crecimiento y convergencia regional en México: aproximación con un modelo panel-espacial,” Estudios Económicos, El Colegio de México, Centro de Estudios Económicos 31, no. 1 (2016): 125-167, https://doi.org/10.24201/ee.v31i1.14.
[24] Jorge Eduardo Mendoza-Cota and Víctor Hugo Torres-Preciado, “The impact of regional remittances on economic growth in Mexico: a dynamic space-time panel approach,” Papeles de Población 25, no. 101 (2019): 113-144, https://doi.org/10.22185/24487147.2019.101.25; González and López, “Remesas, crecimiento y convergencia regional en Mexico.”
[25] Pia M. Orrenius, Madeline Zavodny, Jesus Cañas, and Roberto Coronado, “Do Remittances Boost Economic Development? Evidence from Mexican States,” Federal Reserve Bank of Dallas, Working Paper 1007, October 2010, https://www.dallasfed.org/~/media/documents/research/papers/2010/wp1007.pdf.
[26] Ernesto López-Córdova, Andrea Tokman R. and Eric A. Verhoogen, “Globalization, Migration, and Development: The Role of Mexican Migrant Remittances,” Economía 6, no. 1 (2005): 217-256, https://www.jstor.org/stable/20065489.
[27] Remittances, estimated by Banco de Mexico, include money orders, personal checks, electronic transfers, cash, and in-kind transactions. Therefore, these estimates include formal and informal remittance channels.
[28] The goal of this paper is to analyze how Mexicans, in municipalities and states, use remittances, and determine if these resources generate economic growth and development. There could be a causality from economic growth to remittances, but that is not the objective of this paper. Therefore this causality will not be analyzed.
[29] Jorge Eduardo Mendoza Cota and Cuauhtémoc Calderón, “Impactos regionales de las remesas en el crecimiento económico de México,” Papeles de Población, no. 50 (December 2006):197-221, http://bit.ly/3gmRXhY.
[30] Mendoza Cota and Calderón, “Impactos regionales de las remesas en el crecimiento económico de México.”
[31] The proportion of FDI with respect to state GDP is defined as: Indicator of FDI = (State FDI) / (State GDP). For the cases of remittances and public investment with respect to state GDP, the formulas are the same, dividing the variable with respect to the state GDP. GDP is in constant 2013 prices.
[32] INEGI (Instituto Nacional de Estadística, Geografía e Informática), “Producto Interno Bruto por Entidad Federativa, Año base 2013,” https://www.inegi.org.mx/app/tabulados/default.aspx?pr=17&vr=6&in=2&tp=20&wr=1&cno=2.
[33] The VACB measures the production value added during the economic activities and it is obtained by subtracting the value of total inputs from total gross output. It represents GDP before taxes. See Francisco de Jesús Corona Villavicencio and Jesús López-Pérez, “Obteniendo indicadores de actividad económica municipal basados en información representativa de los Censos Económicos,” Realidad, Datos y Espacio Revista International de Estadística y Geografía 10, no. 2 (August 2019): 62-81, https://rde.inegi.org.mx/index.php/2019/08/20/obteniendo-indicadores-de-actividad-economica-municipal-basados-en-informacion-representativa-de-los-censos-economicos/.
[34] INEGI, “Estadística de finanzas públicas estatales y municipales,” https://www.inegi.org.mx/sistemas/olap/proyectos/bd/continuas/finanzaspublicas/fpest.asp?s=est&c=11288&proy=efipem_fest.
[35] “Informe de Desarrollo Humano Municipal 2010-2015. Transformando México desde lo local,” Programa de las Naciones Unidas para el Desarrollo (PNUD), May 2019, https://www.mx.undp.org/content/mexico/es/home/library/poverty/informe-de-desarrollo-humano-municipal-2010-2015--transformando-.html; “Índice de Desarrollo Humano para las entidades federativas, México 2015,” Programa de las Naciones Unidas para el Desarrollo (PNUD), March 2015, https://www.mx.undp.org/content/mexico/es/home/library/poverty/indice-de-desarrollo-humano-para-las-entidades-federativas--mexi.html; “Human Development Indices,” Global Data Lab, https://globaldatalab.org/shdi/shdi/MEX/?levels=1&interpolation=0&extrapolation=0&nearest_real=0.
[36] Rodolfo de la Torre, “Ten Years of the Human Development Index in Mexico,” Realidad, Datos y Espacio Revista International de Estadística y Geografía 3, no. 3 (September/December 2012):148-163, https://rde.inegi.org.mx/RDE_07/Doctos/RDE_07_Art11.pdf.
[37] CONAPO (Consejo Nacional de Población), “Marginalization Indexes,” Gobierno de México, http://www.conapo.gob.mx/es/CONAPO/Indices_de_Marginacion_Publicaciones.
[38] CONAPO, “Índice de marginación por entidad federativa y municipio 2020: Nota técnico-metodológica,” Secretaría de Gobernación, May 2020, https://www.gob.mx/cms/uploads/attachment/file/634902/Nota_t_cnica_marginaci_n_2020.pdf. This index has five categories of marginalization: very low (-1.52944, -1.15143), low (-1.15143, -0.39539), medium (-0.39539, -0.01738), high (-0.01738, 0.73866), and very high (0.73866, 2.25073).
[39] CONEVAL (Consejo Nacional de Evaluación de la Política de Desarrollo Social), “Información de pobreza y evaluación en las entidades federativas y municipios,” https://www.coneval.org.mx/coordinacion/entidades/Paginas/inicioent.aspx.
[40] INEGI, “Censo de Población y Vivienda 2010,” https://www.inegi.org.mx/programas/ccpv/2010/#Microdatos; CONEVAL, “Mapas de desigualdad 2000-2005,” Secretaría de Gobernación, May 2020, https://www.coneval.org.mx/Medicion/EDP/MP/Paginas/Mapas-de-desigualdad-2000-2005.aspx.
[41] Sistema de Información Económica, “Ingresos por remesas, distribución por municipio - (CE166),” Banco de México,https://www.banxico.org.mx/SieInternet/consultarDirectorioInternetAction.do?sector=1&accion=consultarCuadro&idCuadro=CE166&locale=es; Sistema de Información Económica, “Ingresos por Remesas Distribución por Entidad Federativa,” Banco de México, https://www.banxico.org.mx/SieInternet/consultarDirectorioInternetAction.do?accion=consultarCuadroAnalitico&idCuadro=CA79.
[42] Gobierno de México, “Información estadística de la Inversión Extranjera Directa,” https://datos.gob.mx/busca/dataset/informacion-estadistica-de-la-inversion-extranjera-directa.
[43] “Finanzas públicas estatales y municipales,” Instituto Nacional de Estadística, Geografía e Informática (INEGI), https://www.inegi.org.mx/programas/finanzas/#Tabulados.
[44] I also use the period from 2013 to 2019 for the states, but the results are the same as for the period 2005 to 2019, so I only show these results in the document.
[45] “División territorial,” Cuéntame de México, http://cuentame.inegi.org.mx/territorio/division/default.aspx?tema=T.
[46] Miriam de Regil, “Algunas entidades con bajo nivel de desarrollo en relación al DF: PNUD,” El Financiero, March 4, 2015, https://www.elfinanciero.com.mx/nacional/algunas-entidades-con-bajo-nivel-de-desarrollo-en-relacion-al-df-pnud/.
[47] Uriel Blanco, “Los 10 municipios de México con peor Desarrollo en ingresos, salud y educación,” El Financiero, May 31, 2019, https://www.elfinanciero.com.mx/nacional/los-10-municipios-con-peor-calidad-de-vida-de-mexico/.
[48] CONAPO, “Índice de marginación por entidad federativa y municipio 2020: Nota técnico-metodológica,” Secretaría de Gobernación, May 2020, https://www.gob.mx/cms/uploads/attachment/file/634902/Nota_t_cnica_marginaci_n_2020.pdf.
[49] CONAPO, “Índice de marginación por entidad federativa y municipio 2020.”
[50] Juan José Li Ng, “México: Remesas crecieron 27.1% en 2021, llegan a nuevo máximo histórico,” BBVA Research, February 1, 2022, https://www.bbvaresearch.com/publicaciones/mexico-remesas-crecieron-271-en-2021-llegan-a-nuevo-maximo-historico/.
[51] Li Ng, “México: Remesas crecieron 27.1% en 2021, llegan a nuevo máximo histórico.”
[52] Li Ng, “México: Remesas crecieron 27.1% en 2021, llegan a nuevo máximo histórico.”
[53] Gobierno de México, “Información estadística de la Inversión Extranjera Directa.”
[54] INEGI, “Finanzas públicas estatales y municipales,” https://www.inegi.org.mx/programas/finanzas/#Tabulados.
[55] I use Stata as the statistical software to perform the data analysis and run the regression models.
[56] Campeche is a unique state since it has an extraordinary income derived from oil activity and is very different from the rest of the Mexican states. Campeche is an outlier, and I will include Campeche as I do not have enough data (n = 32), but it does not change the results regarding remittances.
[57] Dummy variables can take any of two quantitative values, usually 1 or 0. Typically, 1 represents the presence of a qualitative attribute, and 0 the absence of that attribute.
[58] Gerardo Esquivel, “Convergencia Regional en México, 1940-1980,” El Trimestre Económico 66, no 264(4) (1999): 725-761, https://www.jstor.org/stable/20857005?seq=1.
[59] Since IM encompasses education, inequality (measured by the GINI coefficient) and poverty conditions, these variables do not have to be included in the econometric model. Indeed, there is a near perfect correlation between IM and those variables in the database.
[60] Francisco de Jesús Corona Villavicencio and Jesús López-Pérez, “Obteniendo indicadores de actividad económica municipal basados en información representativa de los Censos Económicos,” Realidad, Datos y Espacio Revista International de Estadística y Geografía 10, no. 2 (August 2019): 62-81, https://rde.inegi.org.mx/index.php/2019/08/20/obteniendo-indicadores-de-actividad-economica-municipal-basados-en-informacion-representativa-de-los-censos-economicos/.
[61] These correlations are significant at the 1% level (p = 0.0000). The same results occur for the econometric models using GDP and VACB in the case of Mexican states.
[62] CONEVAL measures the percentage of the poor population in the states and municipalities of Mexico. See CONEVAL, “Información de pobreza y evaluación en las entidades federativas y municipios, https://www.coneval.org.mx/coordinacion/entidades/Paginas/inicioent.aspx.
[63] The results for this case are not shown, since this paper focuses only on the variables related to remittances, and the other variables have the same results as before.
This publication was made possible by a grant from the Puentes Consortium.
This material may be quoted or reproduced without prior permission, provided appropriate credit is given to the author and Rice University’s Baker Institute for Public Policy. The views expressed herein are those of the individual author(s), and do not necessarily represent the views of Rice University’s Baker Institute for Public Policy.
