We study firms' responses to two US Environmental Protection Agency (EPA) information-based interventions. First, the EPA disclosed toxicity information on the chemicals listed in the Toxics Release Inventory (TRI). Second, it grouped 17 of the TRI chemicals in the 33/50 voluntary program and challenged firms participating in this program to aggressively reduce their aggregate emissions. Firms therefore faced "twin" signals: focus on the most toxic chemicals, and focus on 33/50 targeted chemicals. We use a novel set of instruments to estimate the causal effects of these twin signals on chemical releases of U.S. manufacturing firms during the life of the 33/50 program (1991-1995), and after the program ended (1996-2013). We examine both "raw" emissions (in pounds) and "weighted" emissions (weighted by toxicity scores) of both 33/50-targeted and non-targeted chemicals. We find that 33/50 program participants reduced weighted emissions of 33/50-targeted chemicals only, with no effects on "raw" emissions or non-targeted chemicals. We also find that these reductions persisted after the program ended in 1995. These results suggest that firms are not unconditional greenwashers or environmental stewards. Rather, firms strategically invest resources to pursue environmental stewardship while taking into account multiple signals from their key sta-keholders such as the EPA.