Pollution haven, green trade agreements and institutional governance: an empirical analysis on a panel of developed and developing countries

Purpose This paper examines whether environmental provisions embedded in trade agreements mitigate the Pollution Haven Hypothesis (PHH), with a particular focus on their effect on outward investment location choices to developing host countries, where evidence remains limited. Design/methodology/approach The analysis relies on a panel of 43 countries over 2009–2021. A gravity model for outward FDI flows is estimated using Poisson Pseudo-Maximum Likelihood (PPML). To address endogeneity and unobserved heterogeneity, the empirical strategy incorporates high-dimensional fixed effects (country-pair, exporter-year and importer-year), complemented by an event-study and placebo tests to detect anticipatory dynamics. Findings Once bilateral heterogeneity and country-specific time-varying shocks are controlled for, the average effect of environmental provisions on outward FDI becomes economically modest and statistically insignificant. Event-study and placebo estimations reveal no evidence of pre-existing trends. Nevertheless, this average conceals substantial institutional heterogeneity. The deterrent effect is primarily concentrated in countries with weak governance and higher corruption, whereas it diminishes and becomes insignificant in better-governed environments. These findings suggest that environmental provisions exert their strongest influence precisely where enforcement capacity is limited, highlighting an institutional dimension of the pollution haven mechanism. Originality/value Rather than treating the PHH as a purely binary phenomenon, this study emphasizes its conditional nature. It shows that the average effect of environmental provisions becomes limited once rigorous fixed-effects are implemented, while institutional quality emerges as a key mediating factor. The paper contributes to the literature on trade policy, environmental regulation and governance in developing countries.