Brander and Lewis argue in a seminal paper (AER, 1986) that a firm's debt-equity ratio should have important strategic effects on product market competition. We test their model in a duopoly experiment under both, Bertrand and Cournot competition. We find that leverage has strategic effects, but those effects are much weaker than predicted by theory. Specifically, we find for price competition a general tendency towards collusion, which has the same overall consequences - but deviates from - the subgame perfect equilibrium prediction. With quantity competition subjects choose much less debt than predicted by theory. It appears that subjects recognize the strategic effects of their own debt. However, they do not (want to) acknowledge possible strategic advantages of opponents' debt.