Working Paper: NBER ID: w24816
Authors: Christopher T. Conlon; Julie Holland Mortimer
Abstract: The diversion ratio for products j and k can be interpreted as the result of an experiment: increase the price of product j, measure the number of consumers who leave product j, and then measure the fraction of leavers who switch to a substitute product k. In theory, this is expressed as the ratio of demand derivatives that enter the multi-product Bertrand-Nash first-order condition for a firm. In practice, diversion ratios are also measured from second-choice data or customer switching surveys. We establish a LATE interpretation of diversion ratios, and show how diversion ratios can be obtained from different interventions (price changes, quality changes, assortment changes) and how those different measures relate to one another and to the underlying properties of demand.
Keywords: diversion ratios; merger analysis; treatment effects; consumer behavior; demand estimation
JEL Codes: L00; L11; L41
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| price increase for product j (L11) | diversion ratio to substitute product k (D10) |
| diversion ratios (L99) | understanding competitive dynamics in differentiated product markets (L13) |
| average diversion ratio (C46) | individual diversion ratios (D30) |
| correctly estimating diversion ratios (C13) | insights into consumer behavior (D12) |