South-South trade: The Gravity Puzzle
Using gravity equations to assess deviations between actual and predicted trade has been the most common approach to study SST. The augmented gravity equation explains the value of bilateral trade by the gross domestic products (GDPs) of exporter and importer, the distance between them, and a vector of control variables accounting for the level of trade resistance between them, such as the level of tariffs, whether they share a common language or a common border, and the existence of a free trade agreement (FTA). In the case of SST, the appealing feature of this approach is that the data involved are readily available.
This approach may be flawed, however, due to omitted variables and endogeneity problems, as examined in Baier and Bergstrand (2006). The latter type of problem is particularly penalizing, since it sheds doubt on the extensive literature addressing the trade impact of integration among developing economies. In a nutshell, are FTAs randomly formed, or are they formed among countries that share characteristics thatmake themgood candidates for integration? If the latter, FTAs’ impact on trade might well be overestimated.
Examining intra-WAEMU (West-Africa Economic and Monetary Union) trade, Coulibaly and Fontagne´ (2006) stress the combined role of economics and geography in the persistence of untapped trade potentials. Being poor and landlocked profoundly reduces trade: beyond distance, the worse combination for SST is accordingly to export from a landlocked country, through a transit country,with a limited share of paved roads.
Poor infrastructure is, however, only a partial explanation for the low level of SST, since it also affects domestic trade. Low levels of SST also result from tariffs, nontariff barriers (NTBs), preferences, and more generally all aspects that make it more difficult to trade with another country than with another region within the same country. Capturing such effects imposes a methodological shift: the right benchmark is domestic trade (the difference between domestic production and exports). A growing literature is using the trade and production database of the United Nations Industrial Development Organization and the World Bank to analyze sectoral trade and production data for a series of countries in the North and in the South. Thismakes it possible to use the so-called border effect methodology to assess whether something specific is hampering SST.
Controllingforsupplycapacity,demand,distance, tariffs, NTBs, and common colonial power, SST is muchmore deflected by the existence of borders than North-North trade orNorth-South trade are.Mayer andZignago (2005) estimate that the impact on SST of the barriers (custom formalities and delays, differences in preferences or regulations, etc.) associated with the existence of a border is the equivalent of a 100 percent tax on imports, and this effect is in addition to that of any existing tariffs and NTBs.
In this context, do free trade areas offer a viable solution? Fontagne´ and Zignago (2007) address this issue in a study designed to control for systematic differences in trade across countries and productive sectors. The six most prominent preferential trade agreements (PTAs) according to their trade impact are the Central American Common Market, the Andean Community, the North American Free Trade Agreement, the European Union, Mercosur (the regional trade agreement among Brazil, Argentina, Paraguay, andUruguay), and the Association of Southeast Asian Nations; other PTAs have a smaller impact on trade.
Another route to trade liberalization among developing countries proceeds from their active participation in the multilateral arena. Simulations conducted with computable general equilibriummodels conclude that this is key to the ‘‘development part’’ of the ongoing multilateral round of negotiations (Francois, vanMeijl, and van Tongeren 2005).
Recent literature reexamines the common view that trade promotes peace and finds that the impact of bilateral and multilateral trade is not the same. Consideringmilitary conflicts in the period between 1950 and 2000, the probability of a conflict is lower among countries tradingmore bilaterally, because of the larger trade losses incurred in case of conflict (Martin,Mayer, andThoenig 2007).Hence regional agreements in the South, as opposed to the multilateral trade liberalization agreements, reinforce the bilateral dependence among developing countries and reduce the probability of occurrence of a conflict.