There is no doubt that migrant remittances (MRs) represent a source of economic development for recipient countries. According to the World Bank (2007), they compensate the lack of public investment. In fact, studies show that MRs have made it possible to fund public infrastructures such as roads, buildings, electricity, drinking water, irrigation systems, mills, factories, dams, etc., social development projects: community centers, literacy projects, schools, dormitories or university dorms, churches and mosques, etc.   . These studies highlight the important role of MRs in funding recipient economies, especially local economies. Sall  expresses this well in this sentence: “In sub-Saharan Africa, all local development is based on migration.” The author explains that the fact that local and regional authorities have no funds of their own has led them to turn towards the benefits of migration. However, it is clear that the actions of migrants for the development of their regions have not significantly impacted the poverty of these regions    . These actions have rather increased poverty in these regions by encouraging the migration of skilled workers  . This raises the issue of how to maximize the impact of MRs on the economic development of recipient countries. In this context, the OECD  suggests several solutions that have been successful in North Africa, Latin America, Asia, Eastern Europe and the Middle East. One notices that these initiatives have not been met with any success in sub-Saharan Africa. The reason for this situation is found in the objectives of African migrants. In fact, the associations of African migrants aim at improving the living conditions of the home village, maintaining the relations with the community and saving face in the community   . In addition, the banking system is not appropriate. In an effort to improve the impacts of MRs on the development of recipient countries, the Chinese and Indian governments have made major political, institutional and tax reforms in order to secure revenues from their migrants  . These solutions can no longer be applied in sub-Saharan African countries because they require huge financial costs that these countries cannot afford  . In addition, private investment initiatives undertaken by migrants in their home areas do not lead to real development for two reasons: on the one hand, the poor business organization due to the lack of operational guidelines and national institutions  ; and secondly, the opportunistic behavior of migrant families   due to the pre-eminence of the group over the individual and the difficulty to innovate. According to Galiègue and Madjimbaye  , this community spirit is the cause for the lack of growth and development of African companies. In other respect, through banks, MRs support micro companies through credit line in Ecuador  . In Brazil, the Inter-American Development Bank and “Banco Sudameris de Investimento” promote entrepreneurship through the creation of a mechanism in which a portion of the MRs is intended for productive investment. In Morocco, too, there is a fund managed by the Banque Populaire  which collects about 60% of all remittances to the country (UNCTAD1  ). This fund provides low interest
loans for real estate investments and entrepreneurs in Morocco. The Unlad organization in the Philippines mobilizes the MRs for productive and community investments. It identifies productive investments and facilitates access to loan. In India, the investment center facilitates business development by working with non-Indian residents to identify sources of capital and technology. In Lesotho, the association of mining workers in South Africa set up a deferred payment plan to ensure that money earned by migrant workers was invested and spent in the domestic economy  . In sub-Saharan Africa, there is a vacuum in terms of solutions for optimizing the impacts of MRs on the development of economies. This study tries to fill this void by proposing to make funds transferred by Burkinabè migrants living in Côte d’Ivoire available to Burkinabe municipalities through loans. To reach its goal, it seeks to answer the following question: “How to transfer private funds over which public authorities have no control to public uses?”
The idea is to analyze the agreement to lend (AL) of Burkinabè migrants to municipalities in their home country. This article is based on a field survey that we conducted in 2014 with 600 Burkinabè migrants living in Côte d’Ivoire. To determine the explanatory factors for agreement to lend, we prefer the Tobit model because we only examine the behavior of people who consent to lend to the municipalities of their country so as to take into account the censored nature of the dependent variable. The rest of the article is presented as follows. We present the situation of decentralization in the first part. The second part explains the methodology and the data used. The third part examines the empirical results obtained.
2. The Actual Situation of Decentralization in Burkina Faso
Decentralization in Burkina Faso, like decentralization in sub-Saharan Africa, faces two challenges: “the weakness of resources and the legitimacy of local authorities. The legality and the legitimacy of the decentralization in Africa remain still ambiguous. Indeed, there are two legitimacies” that of the universal suffrage and that of the usual and/or traditional succession. The mayor is obliged to cooperate with traditional village chiefs      . For Hagberg  , the mayor has the legality and the chief of the village has the legitimacy. The populations recognize themselves much more in traditional chiefs because they are the agents of the custom. As such, they have a stronger influence both at the level of the decision and of the collection of the taxes   . This power of the traditional chiefs also extends to the level of the State. Indeed, in spite of the law on the land tax, the State of Burkina Faso asks for the agreement of the traditional chief for the exploitation of natural resources because the chief has authority on the owner of the land  . According to ESF  , tax revenues are more efficiently recovered by traditional chiefs because of social pressure resulting from community spirit and the funding of the local elected officials depends on the relations between the two entities.
The financing of the decentralization in the countries of sub-Saharan Africa is a real challenge. Indeed, in these countries, only 5% of the state resources are assigned to regions against 10% for all the developed countries  . According to ESF, the resources of the African communities reached 8.5 billion euros of which 7 billion is for South Africa (82%), 1 billion for Morocco and Tunisia is (12%) and 500 million for sub-Saharan Africa (6%). In the Sahelian countries, the rate of local tax collection is estimated on average at 50%. In Burkina Faso, the fiscal pressure is low. It amounts to 7.4% in 2011 against a minimal standard of 17% fixed by the UEMOA  . This weakness is even more stressed at the level of the local finances. According to Chambas and al  , the local tax revenue in Burkina Faso amount to 0.6% of the GDP. Furthermore, the mobilization of own resources extremely turns out to be difficult  . Besides, grants and subsidies of the state are failing taking into account the state’s dependence on the outside world, which is facing an unprecedented economic crisis. For that reason, 80% of the budget of the State of Burkina Faso is externally funded  . This limitation of local resources explains why the social services of Burkina Faso are among the lowest in the world and the resignation of the local authorities from their responsibilities. Moreover, the social pressure also influences the attitude of the migrants leading them to replace local authorities. Migrants are forced to participate in addressing the imbalances in the country  under the penalty of banishment. Therefore, they transfer funds that are used to provide drinking water, electricity, schools, hospitals, police stations, modern equipments for agriculture, subdivision of land etc.  . In other words, Burkinabè migrants are the regulators of the well-being of the populations in their home country.
In the end, it is essential to integrate the cultural aspect into the decentralization process in order to allow local authorities to function   . To increase the impact of MRs in African economies, it is necessary to take into account culture as a fundamental element.
Our study is based on data from a 2014 survey of Burkinabè migrants in Côte d’Ivoire. Since there was no survey frame, building a representative sample with probabilistic methods is impossible. Among the most common non-probabilistic methods is the snowball method, which is appropriate for the study of hard-to-reach populations  . The results of this survey are only about Burkinabè migrants living in Côte d’Ivoire.
2.1. The “Snowball” Sampling Method
Studies dealing with migrants are confronted with a number of obstacles. Indeed, there is no database on migrants in the host countries of the migrants and it is very costly to carry out a survey on migrants in the cases where host countries are many  . Therefore, it is difficult to obtain a sampling frame on the migrants from Burkina Faso living in Cote d’Ivoire. In this case, the probability methods of sampling are impossible. Several methods exist for conducting a field survey on the population hidden or difficult to join. These methods are the methods by chaining or still by networks (Chain-Referral Method) because the population difficult to rich is organized in networks. Among these methods are the Network Sampling, the Adaptive Cluster Sampling, the Snowball Sampling and more recently the Responent-Driven Sampling. Considering the non-availability of a sampling frame on the burkinabés’ migrants in Côte d’Ivoire, the sampling frame method used is the Snowball Sampling. This method is the most used of the non-probability sampling methods. The “snowball” method involves interviewing selected individuals in the target population who serve as initial contacts (first wave). The individuals of the second wave are chosen by those of the first wave. And the process continues until we reach the target number of individuals. However, this is subject to all the biases inherent in chaining methods  . These are the bias of the starting sample that is passed on to the final population, the bias of voluntarism, the bias of concealment and the bias of overrepresentation. The simple random initial sampling in migrant associations allowed us to avoid biases related to the starting sample. We chose five different survey areas in the Soubré2 region, which includes the largest majority of Burkinabè migrants according to the 1998 census. The five survey areas are Grand-Zatry, Mayo, Méagui, Petit Ouaga and Yabayo. For the bias of voluntarism we avoided volunteers by choosing random individuals from associations. Besides, our population is relatively homogeneous concerning the standard of living and the small number of educated people (see Table 1). Regarding the concealment bias, the migrants we have taken into consideration are individuals whom society has no prejudice against. Therefore the migrant has no reason to protect his friends by not mentioning them. So this bias is non-existent in our case study. The same is true for the overrepresentation bias  because our sample consists of a predominantly illiterate population. As a result we have limited the number of individuals recruited to 1 and the number of chains to 6 following  . To do this, we asked each respondent to provide us with the names of three acquaintances; when one of the three people mentioned agrees to answer the questions, the other two are no longer interviewed. The response rate of 88.67% of our field survey indicates that these biases are much lower in our case. On a total of 600 investigated, we record a rate of answer of 88.67%. This rate indicates that these biases are much lower in our case.
2.2. The Selected Model
In order to determine the explanatory factors for the loan consent of Burkinabè migrants, we prefer the Tobit model because we only examine the results from people who agree to lend money to the municipalities of their country. Therefore, it is important to consider the censored nature of the dependent variable since there are migrants who do not consent to lend money. The loan consent variable (LC) of individuals is censored on the left (LC > 0). Besides, we assume that the loan decisions and the choice of the amount of the loan are made simultaneously.
We define our LC as follows:
Y* is the dependent variable that captures the LC of the individual i; Xi is the vector of the independent variables that explains the LC of the individual i, β is a vector of k fixed coefficients to be estimated and εi is a vector of independent residuals normally distributed of zero mean and variance σ². So:
However, the Tobit model estimators are biased in the presence of heteroscedasticity of residues. To fix this problem, we use the Breusch-Pagan Test to test the homoscedascity of our regression.
To evaluate the variables that explain the agreement to lend of Burkinabè migrants, we assume that:
The gender variable has an influence on the agreement to lend. Based on the economic role of women in African culture, we expect the gender variable to have a positive sign. The “Age” variable: the responsibilities assigned to individuals increase with age, something that places elderly people at the top of the hierarchy. Elderly people are more willing than young people to lend to municipalities because they have more responsibilities than young folks regarding the African family. The correlation between age and willingness to lend is positive. The “matsit” variable: the union between man and woman in Africa implies the union of two families and consequently a larger family for man in terms financial responsibilities. Therefore we expect married men to be less interested in lending to municipalities in their home country, which implies a negative sign of the “matsit” variable. The variable “Level of education”: the concept of socio-economic development is better understood by people with a high level of education. The more people are educated, the better they understand the importance of everyone participation in development and poverty reduction. We expect this variable to positively influence our dependent variable. The “profit” variable: migrants lend to municipalities in order to make profits, that is, to increase their earnings. In fact, this is a variable that measures personal interest. The more the migrant agrees to lend, the more he earns profits. Profit increases the willingness to lend. The variable “I love my country”: it represents the bonds of affection that exist between migrants and their country as well as the relations that exist between migrants and their families. This love for the country expresses both a sense of belonging and a sense of attachment to the country. There is therefore a positive relationship between the “I love my country” variable and the willingness to lend. The “By unity” variable: expresses the obligations of unity that comes from the community-based spirit of African culture. The migrant feels obliged to help the weakest for the well-being of the community. The relationship between this variable and the willingness to lend is positive. The variable “Helping the family”: expresses altruism, that is, the migrant’s concern for the well-being of his family. This is therefore a variable that has a positive influence on the willingness to lend. The variable “Citizen”: indicates the active contribution of the migrant in the national public life and its peaceful management. To feel as citizen of the country encourages the migrant to act for the improvement of living conditions in the country. This variable has a positive influence on the willingness to lend. The variable “Reduce Poverty”: measures the participation of migrants in reducing poverty in the country, which means enabling people to meet the basic needs of their own population. We assume that this variable has a positive influence on the willingness to lend. The variable “Express my success”: expresses the migrant’s desire to raise his profile, upon return, he shows everyone that he has taken away the shame on his family. In other words, the migrant saves the family’s honor. Therefore he makes lavish expenses in plain view of the entire community. By contributing to the loan consent, the migrant occasionally shows his success the whole country. This variable is positively related to the willingness to lend. The variable “Job creation”: unemployment is the first reason for immigration of Burkinabè. This variable measures the ability of the migrant to reduce unemployment in his country. In that way, he reduces the migration flows to neighboring countries and outside Africa in order to maintain workforce active in the country. We expect this variable to increase loans to municipalities. The “helping others” variable: expresses support to the weakest, mutual support also comes from the community spirit. This variable positively influences the willingness to lend. The variable “Blessing of God”: the belief in a supreme being is also a feature of African culture. The actions people do in their community aims at attracting God’s blessings in their lives. By seeking to attract God’s blessing, the migrant increases his willingness to lend because he is doing a good work. The variable “All that I have belongs to this country and what belongs to the country is mine” measures the community spirit that implies the responsibility to build a caring, sustainable and right response for the community (the country). The relationship between the willingness to lend and this variable is positive. The variable “economic development”: measures the participation of the migrant in the economic development of his country. That means these actions are in favor of improving the socio-economic situation of the country. This variable has a positive influence on the agreement to lend. By integrating our variables in the Equation (1) we get:
Matsit represents the marital status; Educ, the level of education; Profit, the interest rate of the migrant; Love, the feeling of attachment and belonging; Sol, unity for the country; Fam, the weight of the family; Cit, citizenship; Poverty, the reduction of poverty; Hon, honor of the migrant and his family; Emp, job creation; Mutsup, support to the weakest; Bless, blessing of God; Comspi, community spirit; and Dev, economic development.
3. Results and Interpretation
3.1. Descriptive Results
As displayed in Table 1, the Burkinabe population in Côte d’Ivoire mainly consists of men and youth with a low level of education. The results are similar to those found by Boutin  in his study on Burkinabè migrants in Côte d'Ivoire. He discovered that 9.2% are women compared to 90.8% men. The average age of migrants is 31 years against 33 years for our study and 64% of migrants never went to school compared to 62.9% in our sample. 64.9% of individuals surveyed plan to return to their home country after an average immigration duration of 17 years.
Table 1. Socio-demographic characteristics of migrants.
Source: survey from the author.
Table 2. The reasons of the agreement to lend.
Source: survey from the authors.
3.2. Econometric Results
Table 2 shows the results of the econometric analysis. The specification test of Linktest, the multicollinearity test and the Breusch-Pagan test associated with the regression are given in order to test the robustness of the model (Table 3).
The econometric results show (Table 4) that 43.74% of the variance is explained and the model is globally significant because the p-value (Prob > F = 0.0134) is
Table 3. Estimation of the determinants of agreement to lend using the Tobit model.
***significant at 1%;**significant at 5%; *significant at 10%.
Table 4. The specification Test of Linktest.
Breusch-Pagan test: Homoscedasticiity, chi2 (1) = 4.57 Prob > chi2 = 0. 3392.
below the 5% threshold. It is possible to state that the explanatory variables taken together have a global effect on the dependent variable. Besides, the VIF values associated with the explanatory variables are less than 5, which mean that the regression does not suffer from multicollinearity. The homoscedasticity of the model is verified with the Breusch-Pagan test. The most P-value (Prob > chi2 = 0.3392) is greater than 0.05, so the model is homoscedastic. The Linktest test shows that our model is correctly specified because the predictor squared coefficient is 0.433 (greater than 0.05). So the coefficient is not significant.
Given the lack of documentation on the agreement to lend, we refer to studies on the intent of migrant remittances and the study by Funkhouser  in Guatemala and El Salvador. Funkhouser shows that the difference between the factors that explain the amount of actual MRs and those that explain the intent to transfer is minimal. In addition, Clark and Drinkwater  believe that there is no difference between the transfer decision and the cash transfers actually made. Based on these studies we compare the determinants of microeconomic MRs with the willingness to lend to municipalities which represents the intention of transfer for the loan purposes.
Contrary to our expectations, “the level of education and economic development” negatively influence the agreement to lend. Several reasons can explain these results. In fact, as far as the level of education is concerned, only 0.8% of migrants have a higher level of education, 6.2% have the secondary level. Even so, the higher the education level is, the less people are subjected to culture. For example, 38.5% of migrants with primary education transfer money because “all that belongs to them belongs to Burkina Faso” compared to 25.5% for migrants with secondary level and 0% for migrants with a higher level of education. Finally, the most educated migrants have less confidence in the country’s political authorities. With regard to economic development, given that the vast majority of migrants are out of school, the notion can be misunderstood by migrants. Further, migrants possibly identify economic development with political authorities that have legality but not legitimacy  . So it is necessary to focus first on education, which is the indispensable factor for a better association of African cultural traits with certain factors of sustainable economic and social development. In addition to that, it is necessary to build a climate of trust between authorities and migrants.
The “age” variable is positive and significant at 10% threshold. The older the Burkinabe migrant, the more he agrees to lend to the municipalities of his country. This result reinforces that of Tasiran and Horner  who find that young people have less willing to transfer. In a study about internal and external migrants in the city of Cape Town, South Africa, Tasiran and Horner observe that age positively affects the intent of migrants to transfer in their home regions or countries. In addition, this result confirms Ogandaga’s study  about the fact that elderly people have bigger responsibilities in African culture.
The “matsit” variable is negative and significant at the 5% threshold according to our assumptions. It expresses the size of the household the migrant is responsible of. This result implies that the more the migrant has fewer expenses, the more he agrees to lend to municipalities. It reinforces Tasiran and Horner’s study  which shows that the size of the family negatively influences the decision to transfer funds.
The “Love” variable which expresses the sense of belonging and deep attachment to the country is positive and significant at 1%. Burkinabè migrants agree to lend money to municipalities because of the patriotic feeling and they contribute as a fulfillment of their duties as citizen of the country. This contribution allows them to maintain their membership rights and their identity. The “unity and mutual support” variables are positive and significant at 1% and 10% respectively. This means that the migrant agrees to lend to municipalities in order to accomplish the requirements of unity and mutual support towards the weakest, in this case the municipalities. These variables confirm Fiske’s statement  that the Burkinabe culture is characterized by unity which defines the responsibilities of Burkinabè individuals towards their community and their efforts to restore imbalances.
The “support to the family” variable represents the altruism. It is positive and significant at 5%. Migrants agree to lend to municipalities in order to help their families. It highlights the pressure that the family exerts on the migrant, reinforcing Schrieder and Knerr’s conclusion  in Cameroon. The authors find that the family back home puts pressure on migrants to send back money. The “expression of my success” variable expresses the honor of the migrant and his family. Migrants want to lend to municipalities in order to show that they have value and their immigration was not a failure. This result supports Zongo’s  statement that Burkinabè migrants immigrate in order to be out of the control of elders, achieve their dreams by adding value to their lives and save the honor of their families by paying the head tax. So he backs up Gubert’s result  in Mali which shows that migrants send money to save the honor of their tribes.
The “God’s blessing” variable means that migrants agree to lend in order to attract God’s blessing. It is positive and significant at 1% threshold. In fact, in religions, being charitable attracts God’s blessing which helps people achieve success in all their projects, actions and achievements as they practice charity. She emphasizes the vertical relationship that the African maintains with deity. The “comspi” variable that represents the community spirit is positive and significant at 5%. Migrants agree to lend to municipalities because all that belongs to them belongs to their country. This requires that they participate in the sustainability and good organization, protection and fulfillment of each member of the community (the country). The most important thing for the African migrant is the sense of belonging to the community. This is not for profit or materialist belongings. These results indicate that culture is a powerful determinant in the intention to lend to Burkinabè municipalities. They relate to the study by Clark and Drinkwater  in the United Kingdom; De Voretz and Vadean  in Canada. In fact, Clark and Drinkwater explain through cultural norms, religious, and traditions the difference of attitude between Indians and Caribbean regarding their willingness to transfer. In their study about Asian migrants, North American migrants, Western European migrants and migrants from Southern and Eastern Europe, De Voretz and Vadean note that Asian migrants consider remittances to their parents and/or friends as a normal act of goodness while other migrants see them as luxury. These results also support Lassiter’s  statement that African culture also extends to the descendants African people in the Diasporas. In fact, 48.8% of Burkinabè migrants were born in Côte d’Ivoire. According to the utilitarian model of Cox et al.  , the behavior of Peruvians regarding private transfer is motivated by reasons of exchange and imperfection of the capital markets. In fact, the authors find that being unemployed or sick increases the possibility of receiving private transfers. In other words, the conduct of private transfer is driven by the vulnerability of the beneficiaries. In the case of our study in Burkina Faso, the variables defining the vulnerability of recipients (lack of citizenship, poverty, lack of jobs, the economic underdevelopment of the municipalities) as well as the variable defining the interest of the donor (profit) do not influence the possibility of receiving private transfers. On the other hand, cultural variables such as respect for age, sense of belonging and attachment, unity, honor of the family, mutual support and community spirit and the weight of the family increase the possibility for the municipality to receive private transfers in form of loans. Therefore, these results are contrary to the utilitarian model of migration, which states that all human action is motivated by individual interest, maximization of profit and incomes. In fact, 62.5% of Burkinabè migrants funded their trip on their own while only 2.8% agree to lend to municipalities for profit. Our results also oppose the theory of the new economy of labor migration. Families or communities pay for the expenses related to the immigration of their relatives in order to diversify income or share risk  . However, we find that the variables that make it possible to diversify the risks for the community or the country in our case do not explain the agreement to lend to municipalities. As a result, Burkinabè political authorities can rely on the country’s cultural features in order to have control on MRs which are private funds and use these funds as incomes for municipalities. On the one hand, there is an issue of implementing a policy more focused on migrants to maintain and reinforce the sense of belonging with the organization of national immigrant days or issuing an award for the best immigrated citizen following the example of countries like India. On the other hand, one must be aware that certain African cultural traits are difficult to connect with private initiative and innovation. However, at the community level, these cultural traits are a real asset for development. Further, the immigrant’s need to expose his success, his prestige or the prestige of the family among Burkinabè immigrants in Côte d’Ivoire could be profitable for the municipality by setting up a public reward system for immigrants who lend money. These benefits could be considered as donations knowing that Burkina Faso is a country characterized by a long donation culture  . However, in the pre-survey that we conducted, the migrants refused to give their transfers, believing that their gift does not make a difference in the situation of poverty in the country. This attitude demonstrates that immigration brings about a break with the donation culture in the country. This remark confirms Zongo’s  findings. He finds that immigration leads to a development of initiative spirit and a denial of fatalism.
This study examined the possibility of migrant remittances as a source of funding for municipalities in their home country through an agreement to lend. To do this, we determined the motivations for an agreement to lend. The results showed that migrant remittances can be used to fund municipalities through loans. Given the lack of a study on the agreement to lend, the results were compared to studies on the determinants of remittances. We relied on the Tobit model, which assumes that transfer decisions and transfers are made simultaneously. The results are in contradiction to the utilitarian theory of Cox et al.  that migrant remittances are determined by altruism, unemployment, insurance, etc. In fact, the agreement to loan of Burkinabè migrants is determined by African cultural traits such as community spirit, requirements of unity and mutual support, conformism, gerontocracy, beliefs. This work was limited to a few cultural traits and to one country like community spirit, requirements of unity and mutual support, conformism, gerontocracy, beliefs and to Burkina Faso. It would be interesting to take it over again by trying to test the impact of other African cultural variables on the willingness to lend and also in other African countries.
1UNCTAD: United Nations Conference on Trade and Development.
2Bouquet C., 2003, “Le poids des étrangers en Côte d'Ivoire”, Annales de Géographie, 112.
 Zongo, M. (2009) L’italian dream: Coté cour. L’impact des transferts financiers des migrants bissa en Italie sur les villages de départ dans la provience du Boulgou au Burkina Faso. Annales de l’Université de Ouagadougou-Série A 008.
 Elie, J., Lieber, M. and Lutringer, C. (2011) Migration et développement: Les politiques de la Chine et de l’Inde à l’égard de leurs communautés d’outre-mer. Energie et développement, 2, 215-230.
 Galiègue, X. and ET Madjimbaye, N. (2006) Le management africain, entre contraintes économiques et contingences culturelles: Résultats d’une enquête à N’Djaména, Tchad. Laboratoire d’Economie d’Orléans, Document de Recherche, N°2006-26.
 Azad, A.K. (2005) Chapter 5 Migrant Workers’ Remittances: A Source of Finance for Micro-Enterprise Development in Bangladesh? Remittances Development Impact and Future Prospect, World Bank Publication, Washington DC, 53-76.
 Mondiale, B. (2007) Migrations internationales-une option de sortie par défaut? Rapport confidentiel dans le cadre du programme Rural Struc, Présenté par l’atelier international. Banque Mondiale de Sciences-Po, 2006/2007.
 Bako-Arifari, N., Dicko, A.B., Doevenspeck, M., Sanou, B.W. and Singer, U. (2004) Financer la décentralisation rurale Taxes et impots à l’échelle locale au Bénin, Burkina Faso et Mali. Série: Décentralisation et gouvernance locale, Bulletin, 357.
 Hagberg, S. (2010) Decentralisation and Citizen Participation in West Africa. Bulletin de l’APAD 31-32|2010. [En ligne]
 Ky, A. (2010) Décentralisation au Burkina Faso: Une approche en économie institutionnelle. Thèse de doctorat, Faculté des sciences économiques et sociales de l’université de Fribourg. [El Ligne]:
 Dafflon, B. (2012). Décentralisation et gouvernances locales: Les finances publiques locales des collectivités territoriales au Burkina Faso.
 BAfD (2012) BURKINA FASO, Perspectives économique en Afrique 2012.
 Chambas, G., Brun, J.F. and Grazios, G.R. (2007) La mobilisation des ressources propres locales en Afrique.
 AfBD (2005) Burkina Faso.
 Schrieder, G. and Knerr, B. (2000) Labour Migration as a Social Security Mechanism for Smallholder Households in Sub-Saharan Africa: The Case of Cameroon. Oxford Development Studies, 28, 223-236.
 Zongo, M. (2003) La diaspora burkinabè en Cote d'Ivoire Trajectoire historique, recomposition des dynamiques migratoires et rapport avec le pays d’origine. Politique africaine, 2, 113-126.
 De Voretz, D.J. and Vadean, F.P. (2008) Cultural Difference in the Remittances Behaviour of Households: Evidence from Canadian Micro Data. University of Kent Department of Economics Discussion Papers KDPE 0814.