Estimating effects of domestic transfers using alternative micro-foundations: how do they affect policies assessments? The case of trade liberalization in Senegal

This paper aims at assessing how domestic urban-rural remittances can mitigate macroeconomic shocks in a developing country, especially in terms of income inequities. In particular, when trade liberalization occurs, it may affect the national income structure and increase regional poverty, following shifts in sectoral trade patterns. As underlined by Cox (1990, 2002), Cox and Jimenez (1998) and Morduch (1995), private transfers can significantly help households to deal with exogenous risk and similar studies also find evidence of an efficient risk sharing between the poorest households thanks to private cash exchange (Deaton, 1997; Townsend, 1995; Jalan and Ravallion, 1997). Then, this paper consists in assessing the mitigating impact of remittances on economic shocks by using a micro-macro CGE framework, in which transfer behavior is micro-founded. This model is applied to the case of Senegal1 . Indeed, Senegal is a developing country which is characterized by regional disparities in terms of employment rates, quali_x001c_cations and poverty. These inequalities are particularly marked between urban and rural areas. Furthermore, the choice of Senegal is justified for working on remiting behavior, since the country is the ground of massive internal migration towards urban areas, especially in Dakar. This is heighten by geography and climate: Senegal is located between desert areas (to the North) and tropical zones (to the South) that implies a long dry season, when agricultural activities are diminished. Consequently, Senegal is characterized by temporary migrations towards urban areas, that should affect economic outcomes and policy analysis, especially regarding trade. Three main challenges follow from the target of this paper: 1- the need to model the labor market as reflecting at the closest a dual-dual economy,2, that means to distinguish urban sectors from rural ones and formal activities from informal ones; 2- treating the theoretical ambiguity of the motivations to remit, since there are many theoretical models; and 3- dealing with the credibility of the data on inter-household transfers and treating householdheterogeneity which is essential and justices the choice of a micro-macro framework. In order to treat the first issue (and part of the third one), households are disaggregated as most as possible, following all available criteria in the all set of Senegalese households surveys, namely by region, milieu of living, marital status and number of children, occupation and degree of qualification. This gives us 265 representative households that allow working in a combined micro-macro simulation framework. Regarding the modeling of the labor market, this CGE-model presents a mechanism which endogenizes labor supply and a labor-market segmentation which distinguish the unskilled from the skilled workers. Besides, this CGE-model take into account the double dichotomy between urban and rural areas and formal and informal sectors. The modeling adopted is inspired from Stifel and Thorbecke (2003), but designed in order to match with our sectoral decomposition (34 sectors in the economy, allocated into formal/informal and urban/rural ones, instead of 4 representative sectors in Stifel and Thorbecke, 2003). 3 The second challenge of this work is identifying the determinants of remittances which is puzzling and controversial within the theoretical literature. In the early 1980s, private income transfers have been increasingly recognized as a key economic fact that affects income distributions of an economy subject to political changes. Two main motivations have been considered to explain the decision to remit: altruism and exchange-motivated decision. Based on these two basic principles, numerous microeconomic models have been developed based successively on altruistic motive (Becker, 1974; Stark, 1985, 1995) and mutual exchange strategy (Cox, 1987; Cox, Eiser et Jimenez 1998), the latter being especially relevant in the case of intergenerational transfers (Laferrére and Wolff, 2001). Other common models rely on strategic game analysis (Stark and Wang, 2002), insurance strategy, moral hazard (Stark and Levhari , 1982; Rozenzweig, 1988; Lambert, 1994) and mixed motives (Lucas and Stark , 1985; Andreoni, 1989; and Cox and al., 199S8). Facing these many theoretical conceptions, it is important to check the robustness of our conclusions by implemented dfferent micro-founded functions of transfers and then compare the outputs after simulating external shocks. By the same way, this procedure allows dealing with the third issue of this paper on data availability and credibility, by using a methodology that will be explained below. The last issue addressed is the credibility of transfer data in household surveys. Reliable national data on bilateral remittances is most often not available or inaccurate. To solve this problem, we calculate bilateral remittances from total amounts of paid and received remittances. Following Ratha and Shaw (2007), we allocate the total remittances received among other households using a weight rule, specific to each micro-founded model that is implemented. We apply a three-step methodology for each model of transfers chosen: the first step consists in estimating bilateral transfers among households and using the theoretical motivations to remit as a distribution rule. The second step aims at implementing specific micro-foundations in our CGE model. Finally, the third step consists in implementing different scenarios of trade liberalization: an Economic Partnership Agreement between Senegal and Europe, a worldwide full trade liberalization and thirdly an increase of protectionism. This three-step procedure allows testing the robustness of the conclusions about the link between trade patterns, regional disparities and domestic transfers. The contribution of this paper is to assess and deepen the link between trade liberalization poverty and income inequalities by taking into account domestic transfers as a potential redistributive system. Many CGE papers have included remittances but all of them have considered private cash exchanges as exogenous or fixed as a proportion of the migrant's income. These CGE applications totally ignore the economic literature on the microeconomic foundations of the motivations to remit. This is especially important if we intend to improve the understanding of economic consequences of trade policy on poverty. That is why the paper is organized as following. A first section introduces some stylized facts to show the importance of remitting behavior. A second section comments the existing literature on remittances modeling in CGE analysis, to show the gaps in previous studies and justify this paper. A third section presents the model, especially the modeling of labor market and micro-founded transfers. Next, available data and needed manipulations are presented and dicussed. At final point, trade policies are simulated and their impact on economic outcomes and poverty are assessed, in comparaison with a model that does not include cash-transfers. For this last section, robustness checks are done, by using different micro-foundations of remitting behavior.