PurposeThis research is to examine the impact of online channels on food stockpile behavior.Design/methodology/approachIn this study, we use bivariate probit models to empirically investigate the impact of online purchasing channels on Chinese urban consumer food hoarding behaviors with random survey samples.FindingsResults show that fresh food e-commerce channels are more likely to be associated with panic stockpile behaviors due to higher likelihood of supply shortages than offline channels with government assistance in logistic management. In contrast, community group buy, another format of e-commerce, appears superior in satisfying the consumer needs and easing the panic buying perception.Practical implicationsIt suggests that online channels may have diverse impacts on consumers' panic stockpiling behaviors during the extreme situations. Online channels need to develop efficient supply chains to be more resilient to extreme situations and the government shall recognize the increasing share of the online channels together with traditional offline channels when implementing supporting policies.Social implicationsWith ever increasing share of online channels, it is imperative in terms of policy implications to understand how would online channels affect hoarding behavior.Originality/valueWe are the first study in online shopping's impact on food stockpile during pandemics using a random sample. Although food stockpile behavior at times of emergency have been investigated in many literature, there are no empirical studies on the impact of online channels on stockpile behaviors under extreme situations. Unlike disasters that immediately impact every entity in supply chains covering producers, vendors, distribution centers and retailers, pandemics did not render supply chains affected immediately, but rather increase consumers' willingness to shop online to avoid virus. Thus, Covid-19 provides a natural experiment to investigate the online channels' impact on stockpile behavior.