ABSTRACT The existing literature on the impact of oil supply shocks on real economic activity after the mid-1980s still fails to reach an agreement. This may be partly due to different identification assumptions and model specifications corresponding to the oil supply shock, and partly due to the oil supply shock identified previously is endogenous to real economic activity. In this paper, based on the proxy structural vector autoregressions (SAVRs), we identify an oil supply shock that is exogenous to real economic activity, and then study the impact of this shock on U.S. real economic activity. We find exogenous oil supply shocks have a very large and persistent impact on U.S. real economic activity after the mid-1980s. Besides, several new possible transmission channels through which exogenous oil supply shocks can influence U.S. real GDP are detected. Specifically, unfavorable exogenous oil supply shocks can lower the consumer sentiment index (CSI) and consumer confidence index (CCI), decrease real net exports of goods and services (NEGS) through the terms of trade (TOT) effect, and reduce real government consumption expenditures and gross investment of state and local (GCEGISL), thus leading to a decline in real GDP. These newly detected transmission channels are very important for policymakers.