The retail practice of charging a fee to stock new products is a relatively new but growing phenomenon. Termed a "slotting allowance", it has attracted considerable scrutiny because of uncertainty about its purposes and consequences. We propose and statistically test several hypotheses to assess the degree of empirical support for each of several extant explanations. Slotting allowances, we find, are charged by relatively large retailers who have an informational advantage over the manufacturer about the likely success of the new product. This result apparently contradicts theorizing about the "informational" content of slotting fees, as well as other pro- and anti-competitive explanations. We also find support for the claim that when retailers fear that manufacturers will not provide post-launch support, they pay relatively high wholesale prices.