In this paper, we analyze the spending decisions of over 1.5 million Americans who vary in their degree of revealed credit constraints. Specifically, we analyze how these Americans spend their income tax refunds, using transaction-level data from a stored-value card product. Card-holders may choose among several tax settlement and loan options, effectively receiving cash as much as 90 days earlier than would have been possible without a settlement product. Those selecting earlier settlement options pay higher fees and interest, therefore revealing the level of credit constraints or impatience. We find that more credit constrained or impatient individuals spend their monies more quickly. The mix of cash and merchant transactions is similar between more and less constrained groups. Finally, the primary merchant uses of refunds are to pay for necessities (grocery stores, gas stations, etc.), and the fraction of the refund spending devoted to these necessities is higher for those with greater revealed credit constraints.