A Causal Inquiry into the Economics of Migration and Trade

Al-Malk, Afnan and Maystadt, Jean-Francois and Navarro Paniagua, Maria (2022) A Causal Inquiry into the Economics of Migration and Trade. PhD thesis, Lancaster University.

Text (2022AfnanPhD)
2022AfnanPhD.pdf - Published Version
Download (3MB)

Abstract

Causality is central to economics (Hoover, 2006), for it is the means to predict effects of new interventions and calculate policy counterfactuals (Heckman, 2008). However, in a complex world where “everything depends on everything else” (Valavanis, 1959, as quoted in Hoover, 2004), how does one go about identifying cause and effect? In some applied sciences, Randomized Control Trials provide the gold standard for this purpose (Noble Prize Committee, 2021). The use of experiments offers a satisfactory way since it gives evidence that is both controlled and reproducible (Gower, 2012). However, many questions cannot be addressed by an experiment, either due to financial, ethical or practical constraints (Noble Prize Committee, 2021). Many philosophers viewed the use of statistical methods to be a substitute for experiments, because it allows the extraction of a repeated pattern from a large collection of data (Morgan, 1990). More specifically, it gives scientists a way to deal with plurality of causes in a nonexperimental context (Morgan, 1990). Since the adoption of statistical techniques in economic modeling, it was evident that randomization is not applicable in a similar manner as is the case of controlled experiments, so regression analysis did not essentially provide a causal interpretation (Wold, 1954). To deal with this, a recent breakthrough has been the adoption of the design-based approach that is “aimed at emulating a randomized experiment to answer a causal question using observational data” (Noble Prize Committee, 2021). With the exception of the first chapter, this thesis applies this methodological tool, in particular; exploits a quasi-experimental variation, to answer causal questions on migration and trade. The first chapter examines the effect of religiosity on employment among migrants. The economics literature has recently been more active in examining the effect of religion on different economic outcomes. Religion is found to be an important determinant of individuals’ preferences. In line with some recent work, we deal with the inherent complexity in disentangling religiosity from culture by adopting the epidemiological approach that focuses on migrants. Although an IV strategy, a design-based approach, has been used in the literature to identify the causal effect of religiosity, we decide against its use due to the limitations of IV in this context. We argue that the epidemiological approach along with a rich set of fixed effects that the survey provides allow us to come as close as possible to identifying a causal effect of religiosity on employment. We use one wave of the European Values Study that gives us a sample of 46 European countries in 2008. Our OLS estimates show a negative effect of religiosity on employment. Robustness checks are carried out that confirm the validity of our estimates. We also look into possible mechanisms that could drive this relationship. However, we do not find evidence that any of the potential variables available in our survey can be the channel through which religiosity drives employment. The second chapter, co-authored with Jean-François Maystadt and Maurizio Zanardi, moves into international trade and exploits a natural experiment to better understand the effect of transportation costs on trade. The negative effect of distance on trade is well-established in the literature. However, the debate continues on whether the observed effect is exclusively due to transportation costs or other omitted variables. We take advantage of the blockade that was imposed on the State of Qatar in 2017 to rule out the endogeneity problem and examine how the resulting rise in air transportation costs affected trade. We employ a gravity model estimated using a Poisson pseudomaximum likelihood estimator, and find an air transportation cost elasticity of trade between -0.3 and -0.5. We provide robustness checks to confirm that our results are not driven by potential contaminating factors. The third chapter, co-authored with Jean-François Maystadt and Maria Navarro Paniagua, attempts to get a better understanding of how an exogenous shock affects remittances received by households in Nepal. The blockade on Qatar in 2017 had its share of negative impact on vulnerable migrants. We exploit three waves of a panel dataset on Nepali households between 2016 and 2018. Following the shock, households who had migrants in Qatar experienced a substantial decline in remittances compared to households whose migrants were in different international destinations. We adopt a difference-in-difference approach and control for pre-embargo characteristics to rule out any confounding factors that could bias our estimates. We also show that the decline in remittances is mainly found amongst the poorest households. This result sheds light on the compounded problem of poverty since it is the poor who seem to suffer most and have least resilience in the face of shocks.