THE DETERMINANTS OF FOOD SECURITY IN A GLOBALIZED WORLD: EVALUATION OF THE EFFECTS OF TRADE LIBERALIZATION WITH A SYNTHETIC CONTROL APPROACH

The food security issue is becoming more and more centre-stage in the political and social debate. Especially after the world food crisis of 2007-2008 and the consequent price spikes, it reached the top of the international agenda. The problems of accessibility, disposability and utilization of food, both in developing and poor countries, now draws the attention of policymakers and researchers. Of course, this dynamic and complex problem is influenced and determined by many ‘macro’ factors, analyzed with a case-study econometric approach. The focus of this thesis is to test whether trade liberalization could have a role on the increasing, or decreasing, of the under five mortality rate (WDI, World Bank), used as a proxy for food security. The statistical sample considered is composed of 80 developing countries, 39 of which have experienced an episode of trade liberalization, between 1960 and 2010. The 41 residuals countries are included as a counterfactual in the econometric approach: the Synthetic Control Method (SCM). What was particularly challenging at the beginning of this investigation concerned the isolation of the effect of commercial liberalization on child mortality, and, as a consequence, on the level of food security in each country. As for the methodological approach, what was defined firstly were the covariates: the variables implemented for the synthetic control. They were chosen as a function of their influence on the outcome variable. Above all, we included the level of development and economic growth, the presence of wars or conflicts, the population growth, the percentage of rural population and food supply (measured as the amount of kilo calories available per capita). As already mentioned above with reference to the statistical and the econometric approach, it was developed by Abadie and Garzeabal (2003; 2010), and called SCM. The underlying hypothesis is that it is possible to construct a weighted combination of potential control countries – the synthetic control – that approximate the most relevant characteristics of the country affected by the intervention. In our study, the ‘treated’ country is the one which experienced a trade reform – ‘treatment’. The situation which happened after the treatment (economic openness) can be estimated by the SCM. The SCM is able to calculate a counterfactual circumstance of the ‘treated’ country, in the absence of trade liberalization. The consequential result of the ‘treatment’ can be verified by looking at the outcome trend of the synthetic control. More precisely, the synthetic control algorithm estimates the missing counterfactual as a weighted average of the outcomes of potential controls. The weights are chosen so that the pretreatment outcome and the covariates of the synthetic control are, on average, very similar to those of the ‘treated’ country. Although this methodology is more flexible and transparent than others, there is a limitation regarding the inferential techniques. This is mainly due to the fact that the number of observations in the control pool and the number of periods covered by the sample are usually quite small in comparative case studies, like this. This is why, following the suggestions of the inventors, we implemented the placebo tests, based on permutation techniques. This means that the synthetic algorithm was sequentially applied to every country in the pool of potential controls and then used to compare the placebo with the baseline results. To sum up, we evaluated the dynamic of the under five mortality rate in a ‘treated’ country (which experienced an episode of trade liberalization in the decades considered), with the situation in a sample of control (the synthetic control), composed by a group of very similar countries to the one ‘treated’, but in which there was not any trade reform. It is important to notice that the synthetic control is chosen taking into account the covariates mentioned above; besides, the weight of each country of the synthetic control is selected in order to minimize the difference with the ‘treated’ country, in the pre-treatment period. With respect to the results, we decided to select a pre-treatment period of twenty years, and a post-treatment period of ten years. This is mainly because after a decade from the openness to trade, the effect of this phenomenon of child mortality tend to be very difficult to isolate and identify. The preliminary results obtained, show that the impacts of trade liberalization on under five mortality rate is, on average, positive. Moreover, in a few developing countries political and economic reforms happened simultaneously; so that, this overlapping may not permit the correct identification of the real effect of commercial openness on the outcome variable.