Ducking Data Collection

Consumers may forego transactions when they anticipate that information
collected from their purchases may be sold on and used to discriminate against
them in later markets. We study the sale of consumer data gathered in one
market to a monopolist in a second market. As a side effect of purchasing
in the first market, consumers reveal to Firm 1 their willingness-to-pay in the
second market. For the benchmark case, when consumers do not know their
tastes precisely in the (undisclosed) second market, the firm in the first market
lowers its price to encourage data collection and fish for information. Although
consumers anticipate deleterious consequences in a second market, consumer
surplus and total surplus is higher than without the information fishing. Allowing
consumers to opt-out worsens market performance. The main driver of
this result is the fact that under either policy, consumers who evade having their
data sold impose a negative externality on all consumers in the second market
(since the second firm anticipates that they will have higher-than-average valuations,
and therefore raises its price). We then show there are countervailing
effects when consumers know their demand (type) in the second (exploitive)
market. If data collection is perfect, it induces a hold-up problem whereby the
second market unravels. Data collection raises profit at the expense of consumer
surplus; but allowing opt-out has no effect. We model imperfect/partial
data collection by assuming that the first firm can only discern the tastes of a
fraction of its customers. Then fishing with opt-out is the best policy for firms
and worst for consumers (vis-a-vis laissez-faire data collection or no collection).
Moreover, laissez-faire data collection benefits consumers (at the expense of the
data-harvesting firm) when the data-collection technology is low-yield. When
consumer values in the second market are either low or high, pricing on the second market is mixed (as high value consumers «hide» just up to the point
that the second firm begins to prefer to price to them exclusively.) Finally, we
draw a distinction between opt-in and opt-out, which can have large welfare
consequences: consumers may collectively be substantially better off when it is
opting in, rather than opting out, that incurs a small nuisance cost.

Keywords: data collection, opt-in, opt-out, price discrimination