We obtain laboratory evidence that suggests that (i) promisors are more likely to keep promises the more they were relied upon by a promisee; (ii) this effect is antic-ipated by promisees, who accordingly strategically overinvest to lock promisors into keeping their promises (.psychological lock-in.); and therefore (iii) legal enforcement can reduce overinvestment as promisors don't need to rely on the extra-legal mechanism of psychological lock-in when they can rely on the legal regime instead. We also conduct a vignette study that suggests that punishing behavior by third parties mirrors promise-keeping behavior by promisors. These results contradict the central prediction of the holdup literature that underinvestment will result in the absence of legal enforcement.