Trade policy and food and nutrition security, Background paper prepared for The State of Agricultural Commodity Markets 2015–16. Rome, FAO.

This work contributes to the empirical debate on the impact of agricultural policies on food and nutrition security (FNS). To this aim, it first summarizes some of the arguments and conceptual issues regarding the relationship between agricultural policies and FNS. Whether agricultural trade liberalization improves food security or not is theoretically ambiguous because trade policy influences food production and consumption at both the national and international level. Any change in the trade regime will have a direct effect on rural and urban incomes, and employment, and through these on income distribution. More generally, the nature and magnitude of the effect depends on a number of factors, including: the pace, sequencing and scope of liberalization; the extent of adaptability of the poor to changing economic conditions; the degree of exposure of the country to food imports; the presence of favourable initial conditions and accompanying measures such as adequate regulatory and export capacity; and the time horizon considered. While there is some consensus on the explanation of the empirical evidence on agricultural protection and taxation, the debate on the impact of these measures on food security is still animated and performing a sound assessment is quite challenging. On the one hand, there are multiple links and interactions between trade and FSN at the individual and macro level. On the other hand, there has to be credible exogenous variation in order to establish a causal relationship between public intervention and resulting food security outcomes, but countries may adopt different policies according to distinctive characteristics and this may drive the results in terms of FNS (self-selection bias). To handle these challenges, this work proposes the application of a non-parametric matching technique, namely the Generalized Propensity Score (GPS). This method properly addresses the identification issue between changes in agricultural policies and changes in the expected levels of FNS by controlling for the likely presence of self-selection bias, i.e. unobserved heterogeneity in adopting different policies that may be related to the FNS outcomes. In contrast to previous GPS applications, this exercise looks at the relation between food supply and the actual agricultural policy distortions by countries characterized by similar treatment probability. This allows a contribution to be made to the literature attempting to create FNS typologies for countries at different levels of development and with different agroclimatic conditions or natural endowments for food production. The work presents some results that provide empirical evidence of a significant impact of agricultural policies on food availability. More specifically, countries supporting the primary sector tend to be better off and this is in contrast with the usual claims made by free trade supporters that any policies would do more harm than good. However, the paper also shows that governments may be tempted to provide “too much of a good thing”, since the highest levels of support are associated with lower levels of performance in terms of food availability. The results of the GPS estimates do not allow normative implications to be drawn. However, the evidence of a positive impact suggests that it is worth deepening the analysis, looking at the actual policies implemented by specific countries