We use survey questions about spending in hypothetical scenarios to investigate features of propensities to consume that are useful for distinguishing between consumption theories. We find that (1) responses to unanticipated gains are vastly heterogeneous (either zero or substantially positive); (2) responses increase in the size of the gain, driven by the extensive margin of spending adjustments; (3) responses to losses are much larger and more widespread than responses to gains; and (4) even those with large responses to gains do not respond to news about future gains. These four findings suggest that limited access to disposable resources, and frictions in adjusting consumption, are important determinants of consumption behaviour. We also find that (5) households do not respond to the offer of a one-year interest-free loan, suggesting that this is not a consequence of short-term credit constraints; and (6) people do cut spending in response to news about future losses, suggesting that neither is this a consequence of myopia. A calibrated precautionary savings model with utility costs of changing consumption, and a sufficient fraction of low-wealth households, can account for these features of propensities to consume on both the extensive and intensive margins.