Working Paper: CEPR ID: DP16601
Abstract: This paper studies how firms use service time to respond to local competition and demand conditions when their prices are uniformly set at the national level. Using comprehensive data collected twice a week over three years from 180 Israeli localities, we first show that online grocers set identical prices in all markets where they operate. In contrast, service times are shorter in more competitive markets, on low-demand/low-cost days of the week, and for deliveries offered by high-priced grocers. Next, we exploit regional and temporal variation in entry decisions to examine how incumbents respond. We find that incumbents reduce service time when facing entry, but only on low-demand/low-utilization days. This reduction begins shortly before entry and is greater in monopolistic markets and when entrants pose a larger threat to the incumbent. Service time falls also in markets that do not experience entry yet are served by a fulfillment center serving markets facing entry. We use the classic newsvendor problem model to explain our findings, and in doing so emphasize the importance of accounting for both competitive and supply-side considerations when analyzing firms' response, particularly when prices are unresponsive.
Keywords: online grocery; service time; newsvendor problem; uniform pricing; entry
JEL Codes: D83; L81; L66
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Entry of a new competitor (L13) | Service time of incumbents (C41) |
| Entry of a new competitor (L13) | Service time improvements (C41) |
| Competition in one market (D41) | Service time in adjacent markets (R39) |
| Entry of a new competitor (L13) | Service time in monopolistic markets (D42) |
| Entry of a new competitor posing a larger threat (L13) | Service time of incumbents (C41) |
| High-demand, high-utilization days (L97) | Service time of incumbents (C41) |