The purposes of this study were to evaluate the presence or absence of pain with the jerk test as a predictor of the success of nonoperative treatment for posteroinferior instability of the shoulder and to identify pathologic lesion responsible for the pain in the jerk test. Eighty-nine shoulders(81 patients), which had posteroinferior instability with positive posterior clunk in the jerk test, were nonoperatively treated. The patients were divided into two groups with respect to the presence of pain in the jerk test: painless jerk group(54 shoulders) and painful jerk group(35 shoulders). Response to the nonoperative treatment was evaluated after at least 6 months rehabilitation program. Patients who did not respond to the rehabilitation underwent arthroscopic examination to identify any pathologic lesion. The painful jerk group had higher failure rate with nonoperative treatment (p<0.001). In the painless jerk group, fifty shoulders (93%) responded to rehabilitation program after a mean of 4 months. Four shoulders(7%) were unresponsive to the rehabilitation. In the painful jerk group, five shoulders(16%) were successful with the rehabilitation while the other thirty shoulders(84%) failed. All 34 shoulders, which were unresponsive to the rehabilitation, had a variable degree of posteroinferior labral lesions. In conclusion, the jerk test is a hallmark for predicting the prognosis of nonoperative treatment in the posteroinferior instability. Shoulders with symptomatic posteroinferior instability and a painful jerk test have posteroinferior labral lesion.