Payments for ecosystem services (and conditional cash transfers more broadly) can align incentives in a flexible manner to increase targeted outcomes. However, incomplete markets like financial or informational frictions, or weak institutions which reduce trust in these programs can reduce compliance and therefore program efficacy. We use a randomized controlled trial to test the efficacy of cash transfers on reducing agricultural residue burning, which causes enormous health externalities due to air pollution. To mitigate these frictions, we randomize paying a portion of the cash transfer up-front (unconditionally). Despite lower marginal payments for compliance, farmers in upfront payment programs reduce burning by over 10 percentage points, while those in the standard PES program show much smaller reductions in burning.