Replication files for "The Generalized System of Preferences and NGO Activism"
Research question: We study if NGO activism can contribute to the implementation of Labor Laws in a developing country. We exploit as a quasi-natural experiment the renegotiation of the Generalized System of Preferences (GSP) between the US and Indonesia in 1994, which induced the Indonesian government to raise the level of the legal minimum wage. We show that the activism of workers’ rights groups helped increase firm-level average wages up to the minimum-wage level, not only inside but also outside the export sector. Labor NGO activism helped to implement the new minimum wage standards in a country that lacked strong governmental institutions. Strategy: We exploit the geographical variation in labor NGO activism and the exogenous shock, the deadline of the US GSP revision. Our first objective is to analyze the impact of labor NGO activism on average firm wages. We consider Districts with active labor NGOs as the “treatment” group and those without as “control” group and carry out a difference-in-difference estimation. We want then to identify whether the presence of labor NGOs shaped the gap between the minimum wage and the firm level wage. The main outcome variable is lnwage gap, the log of the difference between the minimum wage and firm-level average wages. Our last identification strategy relies on an event study, in order to confirm the absence of differential trends in districts with labor NGOs. The omitted category is the year = 1994, the year of the deadline of the US GSP revision. We perform event study analysis for the following dependent variables: i) the log-wage, ii) the wage gap, defined as the difference (in levels) between the minimum wage and average wage at the firm level and iii) violation, defined as the wage gap over the minimum wage. Data: Our main source of firm data is the annual Manufacturing Survey of Indonesia, collected and compiled by the Indonesian government statistical agency.We have observations on around 20,000 manufacturing firms per year. The dataset covers the period before and after the US GSP revision in Indonesia: 1991-1996. The second source of information covers labor NGO activism in Indonesia, and relies on a detailed list of labor NGOs compiled by Ford (2009). We use information on labor NGO location at the District level. In the robustness analysis we use information on the District-level location of other Indonesian humanitarian NGOs, publicly available in the dataset on the website of SMERU, a large Indonesian NGO. Dofile: The dofiles replicate tables and figures in the paper.