This article evaluates the short-term impacts of a fruit fly integrated pest management program in Peru. Exploiting arbitrary variation in the program's intervention borders, we use a geographical regression discontinuity design to identify the program's effects on agricultural outcomes. Pre-treatment balance tests show that producer and farm-level pre-treatment characteristics evolve smoothly at the intervention border. Results indicate that farmers within treated areas improved pest knowledge and are more likely to implement prevention and control practices. Also, they increased fruit production and sales. Our findings are confirmed by placebo tests and are robust to alternative regression discontinuity bandwidths and polynomials.