The control of pests by their natural enemies represents an important ecosystem service that maintains the stability of agroecosystems and has the potential to mitigate pest control costs both to private producers and to society. Extending the "economic threshold" concept, this paper proposes an "ecological economic threshold" for pesticide use that takes into account the implicit cost of injury to natural enemies. By explicitly accounting for natural pest suppression, the ecological economic threshold can potentially make pest management more cost-effective while reducing dependence on toxic insecticides. The threshold is illustrated via an intra-seasonal dynamic bioeconomic model of soybean aphid management in Michigan, USA. A dynamic programming model quantifies the economic value of natural suppression to optimal pest control. The results highlight the importance of assessing both pest and natural enemy populations in making insecticide application decisions and accounting for the opportunity cost of insecticide use due to its collateral damage effect on natural enemies. The paper offers a preliminary, lower bound estimate of the value of natural pest predation as inferred from the insecticide input replacement cost per acre for the area where the natural enemy complex can suppress pest population below the level at which it causes damage. A sensitivity analysis shows that numerical solutions for the ecological economic threshold are sensitive to biological parameters such as predation rate and net growth rate of pest population. We recommend that future research move beyond insecticide thresholds to develop guidelines for explicit management of habitat for the natural enemies of agricultural pests.