How free trade can boost welfare

Once known as the “Safe Haven” of Western Africa, because of its long-standing political stability and economic success, Côte d’Ivoire plunged in a decade-long vicious circle of political violence after a coup d’état in December 1999. The level and scope of violence reached its peak in September 2002 when a coalition of three rebel movements, known as the Forces Nouvelles de Côte d’Ivoire (hereafter FNCI), occupied and tightened its grip over 60% of the country’s territory. Unlike other rebel movements in West African states such as Liberia and Sierra Leone, where territorial conquests were allegedly associated with “scorched-earth” and “denial-of-resource” tactics, the FNCI opted for an autonomous self-governance system.

I was in my sophomore year at the University of Bouaké when the rebels stormed the town and established it as their headquarters. They preserved some of the public infrastructures including power plants and water systems. Small businesses like bakeries, convenience stores, restaurants, and even several factories continued to operate. Local and international NGOs were also allowed to provide a minimum service in education and health (Balint-Kurti, 2007). To finance operations, the FNCI set up a sophisticated tax and customs organization, known as La Centrale, which levied export tax on cocoa beans in the jurisdictions they controlled (Guesnet et al., 2010). Côte d’Ivoire is the world’s largest supplier of cocoa and the FNCI’s zone was estimated to comprise 10-25% of the nation’s annual production of 1.26 million metric tons (Witness, 2007).

Relying on agricultural export taxes to raise revenue is not uncommon in developing countries. In a context where tax administrations are often dysfunctional and ineffective, export taxation has become an attractive policy instrument for many governments. Such an easy-to-implement trade barrier functions as a shield against revenue losses due to declining prices and deteriorating terms of trade in agricultural commodities. Economists have warned that high and persistent agricultural export tax rates can depress farmers’ earnings, deter production, and eventually decrease public receipts (McMillan, 2001). Yet, there is little if any empirical evidence of a causative link between agricultural export taxation and farmers’ living standards. My job market paper (JMP), titled Export Taxes and Consumption: `A Natural Experiment’ from Côte d’Ivoire, fills this gap by exploiting the advent of two de facto states, during the Ivorian civil conflict of 2002-2007, to examine how export taxation affects the life of rural peasants.

Before the partition of Côte d’Ivoire into two self-governed entities, a uniform export tax scheme prevailed throughout the country. Dichotomous export taxes only surfaced with the apparition of the FNCI in 2002. While the Droit Unique de Sortie (DUS) – the main export tax on cocoa beans – was maintained at its pre-partition rate of 220 FCFA ($0.44) per kilogram in the government-administered regions, its equivalent in the rebel-held provinces fluctuated between 50 FCFA ($0.10) and 150 FCFA ($0.30) per kilogram (Witness, 2007). Did these large reductions in export tariffs actually translate into tangible benefits for cocoa farmers in the rebel-held north? Did the conflict, a deleterious political event by definition, prevent the farmers from benefiting from the liberalization policy?

During my doctoral studies at the University of Oklahoma, I decided to investigate these issues more systematically. I gathered the pre (2002) and post (2008) partition waves of the household survey “Enquête sur le Niveau des Vie des Ménages” from the National Institute of Statistics of Côte d’Ivoire. The data contain a rich set of information about living standards, demographics, location, etc. of 23,400 households, among which 4,891 individuals are identified as cocoa farmers. At first glance, the raw data revealed that non-cocoa farmers, those who were not directly affected by the policy, showed no significant changes in their expenditure patterns (see upper panel of the expenditure distribution displayed in the figure). In contrast, cocoa farmers in the rebel-held north, the beneficiaries of the tariff reduction, experienced a dramatic catch up in living standards relative to their southern counterparts (lower panel).

To further examine the effects of this change causally, I employed a more robust econometric strategy in the form of the difference-in-difference-in-differences approach. Specifically, I compared changes in expenditures between cocoa and non-cocoa farmers in the rebel-held state (treatment state) with changes in expenditures between cocoa and non-cocoa farmers in the government-controlled state (control state), before and after the division. I augmented the model with a rich set of individual and household levels controls while accounting for time invariant province-specific characteristics. To refine the identification, I also use a sub-sample of communities that resided in districts split by the demarcation line separating the two states. Because district creation often obeys certain socio-economic, cultural and geographic coherences, treated and control groups from these split districts would have the advantage of being more comparable.

My baseline specification suggests that cocoa farmers in low export tax jurisdictions experienced a significant increase in total consumption expenditure of about 43% relative to their peers in high export tax provinces. This was equivalent to an increase in consumption from the sample mean of 88,740 FCFA ($177.5) to approximately 126,900 FCFA ($254). Furthermore, the impact of the policy was not homogeneous across income groups or across categories of expenditure. In fact, quantile analysis suggests that the policy had the biggest effect on the poorest cocoa farmers. In terms of categories of expenditures, affected farmers allocated relatively more resources to their basic needs such as food, clothing, and healthcare. However, relative investments in children’s education did not significantly change following the implementation of the preferential export tax policy. One possibility is that this could reflect the provision of cost-saving incentives to parents by the NGOs that run schools in the rebel-held areas (Sany, 2010). An alternative is that favorable conditions in the cocoa sector and anticipation by farmers of a long-lasting beneficial policy may have pulled children away from schools and into the cocoa fields (Nkamleu and Kielland, 2006).

Finally, using information on farm-gate prices from the surveys, I show that cocoa farmers in the North received on average approximately 60 FCFA ($0.12) more than farmers in the South for each kilogram of cocoa beans sold. This suggests that the transmission of border prices to local producers is a relevant channel through which dismantling agricultural export barriers boosts farmers’ earnings.

To sum up, I exploit this rare opportunity of a “natural experiment,” characterized by the temporary coexistence of two de facto “states” in Côte d’Ivoire to show that agricultural export restrictions can perpetuate poverty in farming households. By establishing a causal link between an agricultural trade policy and poverty, my JMP contributes to a much broader debate on the economic consequences of freer trade in developing countries (for a review see Winters et al., 2004).

This post first appeared on The World Bank

Author: Souleymane Soumahoro is a PhD Candidate in Economics from the University of Oklahoma, and an affiliate of Household in Conflict Network (HiCN).

Image: A farmer holds organic coffee beans. REUTERS/Enrique Castro-Mendivil