We offer experimental and theoretical evidence that the auction method for initial public offerings (IPOs) may be improved through the use of hybrid auctions with separate retail tranches or ‘public pools’. Such hybrids, which combine a price-setting tranche (an auction or book building) with a separate tranche that allows investors to place orders without specifying a price (the public pool), have been used for IPOs around the world. We develop theory, then run laboratory experiments to examine the effects of a public pool on multi-unit uniform price auction IPOs. Our experimental auction design incorporates key features of the IPO process such as endogenous bidder entry, costly information acquisition, differing capacity constraints and uncertainty with respect to the intrinsic value. Simulations are used to characterize the Symmetric Bayesian Nash Equilibria (SBNE) for both pure and hybrid auctions in a model that is calibrated to key parameters from our experimental data, generating predictions for the remaining variables. As predicted, a public pool tranche improves auction performance by increasing proceeds, lowering price volatility, reducing price error and reducing the incentive for small bidders to free ride by submitting extremely high bids. Underpricing occurs in both treatments but is less severe with the public pool. We also show that in collusive-seeming multi-unit auction equilibria, it may be optimal for informed, rational bidders to place clinching bids strictly above the expected value per unit, leading to very steep demand curves. Overall, our results imply that both IPO auctions and crowdfunding may be improved by restricting retail investors to a separate, non-price-setting tranche.