Auctions often involve goods exhibiting a common knowledge ex-post risk that is independent of buyers’ private values or their signals regarding common value components. Esö and White (2004) showed theoretically that ex-post risk leads to precautionary bidding for DARA bidders: Agents reduce their bids by more than their appropriate risk premium. Testing precautionary bidding with data from the field seems almost impossible. We conduct experimental first-price auctions that allow us to directly identify the precautionary premium and find clear evidence for precautionary bidding. Bidders are significantly better off when a risky object rather than an equally valued sure object is auctioned. Our results are robust if we control for potentially confounding decision biases.