Why do states overestimate threats and, as a result, mount disproportionately strong and therefore costly balancing responses? To answer this question, we build a neoclassical realist theory of overbalancing to argue that unit-level intervening variables help generate a counterforce greater than what a structurally induced ideal response would call for. We identify the factors and conditions that steer states to deviate from realist, optimal policies, pinpoint the consequences of such suboptimal behavior , and provide policymakers with recommendations more suited to an interest-driven foreign policy in line with power considerations. We apply our theory to two distinct case studies: Egypt's costly intervention in Yemen in the 1960s and the American overreaction to the real, but very limited, threat posed by terrorism since 2001.