The purpose of this study was to develop a fracture tolerance for the elbow joint, or proximal ends of the ulna and radius, relative to the fracture risk under side-impact airbag loading. Forty experiments were performed on the elbow joints of small female cadavers. The energy source, a pneumatic impactor, was configured to apply compressive loads that match the onset rate, peak force, and momentum transfer of previously conducted side-impact airbag tests with small female subjects. Three initial orientations of the impact load angle relative to the longitudinal axis of the forearm were selected based on analysis of side-impact airbag tests with the instrumented dummy upper extremity. These included loading directions that are 0°, 20°, and 30° superior of the longitudinal axis of the forearm. Post-test necropsy revealed that 11 of the 40 tests resulted in chondral, osteochondral, or comminuted fractures of the proximal radial head or the distal trochlear notch. Using the fracture outcome as the binary variable, a generalized estimating equations statistical analysis showed a significant correlation between elbow load angle (p < 0.01) and risk of fracture, as well as peak elbow force (p = 0.04) and risk of fracture. Using data that were mass scaled to the 5th percentile female, the analysis produced a multivariate fracture risk function that predicts a 50% risk of elbow fracture at a compressive elbow load of 1780 N and load angle of 30° superior to the longitudinal axis of the forearm (p < 0.01). It is anticipated that this tolerance will be used to reduce the risk of elbow fractures from side airbag deployment.