AbstractI use a quasi experiment, conducted in the field, to evaluate how the workers of an economic firm responded to threats to cut their piece rates. During the study, the piece rate paid to workers was initially set above the regular rate, given working conditions. In a commitment treatment this high rate was paid without any threat of revision. During a non‐commitment trial, the high rate was accompanied by a threat from management to reduce the rate if, after two days of work, average earnings were too high. The empirical results display strong ratchet effects. Workers withheld output by 16% under non‐commitment relative to commitment. The firm eventually reduced the piece rate from the initial high rate, but left it above the regular rate, giving workers the possibility to extract rents. Workers showed no tendency to restrict output during this phase of the study.