DP13793 Price Discrimination in the Information Age: Prices, Poaching, and Privacy with Personalized Targeted Discounts
|Author(s):||Simon P Anderson, Alicia Baik, Nathan Larson|
|Publication Date:||June 2019|
|Keyword(s):||competitive price discrimination, discounting, GDPR, opt-in, privacy, targeted advertising|
|JEL(s):||D43, L12, L13, M37|
|Programme Areas:||Industrial Organization|
|Link to this Page:||cepr.org/active/publications/discussion_papers/dp.php?dpno=13793|
We study list price competition when firms can individually target discounts (at a cost) to consumers afterwards, and we address recent regulation (such as the GDPR in Europe) that has empowered consumers to protect their privacy by allowing them to choose whether to opt in to data-gathering and targeting. In equilibrium, consumers who can be targeted receive poaching and retention discount offers from their top two firms. These offers are in mixed strategies, but final profits on such a consumer are simple and Bertrand-like. More contestable consumers receive more ads and are more likely to buy the wrong product. Poaching exceeds retention when targeting is expensive, but this reverses when targeting is cheap. Absent opt-in choice, firm list pricing resembles monopoly, as marginal consumers are lost to the lowest feasible poaching offer, not to another firm's list price. Opt-in choice reintroduces the standard margin too on those who opt out. The winners and losers when targeting is unrestricted (rather than banned) depend on the curvature of demand. For the empirically plausible case (convex but log-concave), targeting pushes up list prices, reduces profits and total welfare, and (if demand is convex enough) hurts consumers on average. Outside of this case, more convex (concave) demand tends to make targeting more advantageous to firms (consumers). We then use our model to study the welfare effects of a policy that forbids targeted advertising to consumers who have not opted in. Consumers opt in or out depending on whether expected discounts outweigh the cost of foregone privacy. For empirically relevant demand structures, allowing opt-in makes all consumers better-off.