Is online retail killing coffee shops? Estimating the winners and losers of online retail using customer transaction microdata

Is online retail a complement or substitute to local offline economies? This paper provides the first evidence that consumers use time saved from online retail to increase their trips for time-intensive services like coffee shops. I use new, detailed data on the daily transactions of millions of anonymized customers. I then estimate a discrete choice model of consumer trip choice, which embeds time use mechanisms and accounts for correlations in trip utility shocks. I show that the model matches key features of observed behaviour that are missed by more standard models, such as the disproportionate increase in trips to nearby coffee shops when consumers switch to online groceries. Model counterfactuals are used to forecast changes in future trip demand and outline strategies, which offline retailers can use to compete against online retail. For consumers, I find that the welfare gains from online grocery platforms go disproportionately to high-income consumers.