What are the expected effects of trade policies on Poverty in Senegal? a CGE Micro-Macro Analysis

There is an ongoing debate on the role of trade policies in alleviating poverty. Indeed trade liberalization is supposed to improve economic growth (Dollar and Kraay, 2002; Irwin and Tervio, 2002; Frankel and Romer, 1999). Focusing on poverty alleviation and income inequalities, the positive impact of trade is less consensual. Some works have defended the idea that trade integration implies poverty reduction (Bhagwati and Srinivasan, 2002; Dollar and Kraay, 2004; Anderson and Martin, 2005), but most recent studies have not reached this general conclusion, pointing that the link between trade and poverty can be puzzling (Winters, McCulloh and McKay, 2004; Hertel and Winters, 2006; Goldberg and Pavcnik, 2007; Harrison, 2007). According to these studies, trade policies bring contrasted effects on poverty but region or sector-specific conclusions can be done. This paper aims at assessing the expected effects of trade policies on poverty reduction in Senegal. Especially, the main issue is to point out the distributional effects of trade policies among households, following regional, sectoral, occupational and skills features. Our study consists in building a single-CGE model, adapted to poor countries and doing counter-factual micro-simulation analysis to underline the income and distributional effects of tariff-reducing under different scenarios. Thus, in order to match with the Senegalese economy, our CGE-model framework arises from two main issues: treating households heterogeneity and modeling the labor market in order to reflect at the closest a dual-dual economy (Stifel and Thorbecke, 2003). This concept refers to the double dichotomy between urban and rural areas and formal and informal sectors. It implies to distinguish urban from rural sectors and formal from informal activities. To treat the first issue, we disaggregate households as most as possible, following all available criteria in the all set of Senegalese households surveys, namely by region and milieu of living, marital status and number of children, occupation and degree of qualification. This gives us 265 representative households that allow us to work in a combined micro-macro simulation framework. By this way, it is possible to develop a model in which different kinds of workers can be modeled and thus address our second issue (namely modeling a dual-dual economy). Indeed, many of the classical CGE studies in international trade work with simple sets of assumptions about the labor market that are not appropriated to developing countries, assuming especially fixed or uniform labor supply. Thus, to address this, our CGE model presents a mechanism which endogenizes labor supply and a labor-market segmentation which distinguish the unskilled from the skilled workers. This allows us to capture the skill-specific labor market effects of shifts in international trade patterns. Besides, the distinction between workers attached to the rural versus the urban sector is important, since regional mobility must be taken into account. Finally, we take into account mobility between formal and informal sectors because productivity and wages differentials imply different effects of trade policies. As in most CGE models, formal and informal labor are used in separate sectors.3 We decide to adopt a modeling that is inspired from Stifel and Thorbecke (2003), but design it 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). As underlined by Boeters and Savard (2011), this kind of modeling brings new issues such as the need to obtain labor supply estimates that can be used in our combined micro-macro model.4 After the model is designed and calibrated on a SAM built for the year 2006 (Fall, 2011), different scenarios of trade policies are applied. The first scenario is an EPA agreement between Senegal and Europe. Indeed, the EU and its ACP partners were unable to conclude the EPA negotiations as planned on January 1st 2008 and this is still an ongoing process. The second scenario is full liberalization. These trade policies have already been assessed in the literature, but mainly using multi-countries CGE models (Berisha-Krasniqi et al., 2008; Fall et al., 2007) or using a dynamic recursive computable general equilibrium (Cissokho and Diop, 2011). Our micro-macro framework is complementary and necessary to evaluate the impact in terms of poverty alleviation.