Dynamic competition in deceptive markets

• Date: 01 June 2020
• Author: Johannes Johnen
• DOI: http://doi.org/10.1111/1756-2171.12318
• Published date: 01 June 2020

RAND Journal of Economics

Vol.51, No. 2, Summer 2020

pp. 375–401

Dynamic competition in deceptive markets

Johannes Johnen∗

In many deceptive markets, firms design contracts to exploit mistakes of naive consumers. These

contracts also attract less-profitable sophisticated consumers. I study such markets when firms

compete repeatedly. By observing their customers’ usage patterns, firms acquire private infor-

mation about their level of naiveté. First, I find that private information on naiveté mitigates

competition and is of great value even with homogeneous products. Second, competition between

initially symmetrically informed firms is mitigated when firms can educate naifs about mistakes.

In an analogous setting without naifs, the second result does not occur; the first result occurs

when firms cannot disclose fees.

1. Introduction

Both intuition and extensive empirical evidence suggest that firms in many markets have

a better understanding of consumer behavior than consumers themselves. This allows firms to

exploit consumer misunderstandings. Examples are markets for credit cards (Ausubel, 1991;

Agarwal et al., 2008; Stango and Zinman, 2009, 2014), retail banking (Alan et al., 2018; Compe-

tition and Markets Authority,2016; Cr uickshank, 2000; Officeof Fair Trading, 2008), or mobile-

phone services (Grubb and Osborne, 2014). In most of these markets, firms and consumers

interact repeatedly. Yet the existing theoretical work—such as Gabaix and Laibson (2006),

Grubb (2009), Armstrong and Vickers (2012), and Heidhues, K˝

oszegi, and Murooka (2016b)—

considers non-repeated interactions.1In these models, some naive consumers are unaware of

shrouded or hidden attributes. I extend this literature by introducing a dynamic model of com-

petition with deceptive products. This allows me to investigate an increasingly relevant aspect

of reality: developments in the analysis of large amounts of data help firms to predict consumer

behavior with increasing precision. By evaluating their own customers’ usage data, firms have

an informational advantage relative to their competitors. I ask how this informational advantage

∗CORE/LIDAM, Université catholique de Louvain,Voie du Roman Pays 34, Louvain-la-Neuve, 1348, Belgium.

I thank David Myatt, three anonymous referees, Alexei Alexandrov, Özlem Bedre-Defolie, Helmut Bester, Yves

Breitmoser, Ulrich Doraszelski, Michael D. Grubb, Paul Heidhues, Johannes Hörner, SteffenHuck, Rajshri Jayaraman,

Botond K˝

oszegi, Dorothea Kübler,Takeshi Murooka, VolkerNocke, Martin Peitz, David Ronayne, Marc Rysman, Heiner

Schumacher, Benjamin Solow, Rani Spiegler, Konrad Stahl, Roland Strausz, GeorgWeizsäcker, participants at the 2015

MaCCI Workshop on Behavioral IO in Bad Homburg, and seminar and conference audiences. I thank the FW-B for

funding the Action de recherche concertée grant n°19/24-101 “PROSEco”, and the FNRS and FWO for the EOS project

30544469 “IWABE”.

1Gabaix and Laibson (2006) consider an extension where consumers buy a base product once, but an add-on

multiple times, but they do not allow for repeated interaction where firms adjust conditions overtime.

© 2020, The RAND Corporation. 375

376 / THE RAND JOURNAL OF ECONOMICS

affects competition when some consumers are more prone to making mistakes than others. I then

compare results to an analogous setting without naive consumers.

Formally, I study a two-period model with shrouded product attributes. I≥2firmssella

homogeneous good in each period. Firms charge a transparent price and a hidden fee, that is,

the shrouded attribute. There are naive and sophisticated consumers. Naifs pay the hidden fee

but do not take it into account. Sophisticates are aware of the hidden fee and avoid it. In this

way, the hidden fee represents unexpected payments due to costly mistakes.2The novel feature

is that firms analyze their customers’ usage patterns to predict their behavior, and to target offers

accordingly: in period 1, firms compete with symmetric information on consumers. After observ-

ing which of their first-period customers paid the hidden fee, firms learn to distinguish between

naifs and sophisticates within their customer base. In period 2, firms use this private information

to discriminate between continuing customers based on their level of sophistication.

Credit cards are an example of a market close to this setting. The market is competitive by

conventional measures, that is, many firms sell a quite homogeneous product. Consumers are

usually aware of maintenance fees, cash rewards, introductory APRs, or new-client bonuses, but

many consumers ignore over-limit fees, late fees, or underestimate their future borrowing when

choosing a credit card. Firms condition offers on customer characteristics, including naiveté,

which they can infer from usage data.

Following Gabaix and Laibson (2006), many previous ar ticles on deceptive products study

a similar setting, but with non-repeated interaction where firms cannot distinguish consumers.

Firms charge large hidden fees, making naifs more profitable than sophisticates who choose the

same contract, but competition with transparent prices drives away any profits. Profits from naifs

end up cross-subsidizing sophisticates.

My first main result is that private information on naiveté creates profits in period 2. Firms

use this information to reduce the intensity of competition. When a firm Aobserves past usage

patterns in her customer base, she learns to distinguish between her old customers in period 2,

but A’s rivals remain uninformed about A’s customer base. Acan use her information to target

continuing naifs with a smaller transparent price than sophisticates. Sophisticates do not get this

transparent discount and are more prone to switch to a rival. Thus, uninformed rivals cannot

attract A’s profitable naifs without also attracting unprofitable sophisticates. Because of this ad-

verse selection of sophisticates, rivals compete less vigorously. By inducing adverse selection,

firms can profitably exploit the fact that private information on naiveté allows them to discrimi-

nate between their continuing customers, whereas rivals lack the information to do so, and naifs

lack the sophistication to recognize better offers.

Firms use naiveté to induce adverse selection. With onlyrational consumers and obser vable

fees, competition severely limits the value of information. I establish this in a benchmark with

only sophisticated consumers. All value a base product and some also value an add-on. Firms use

private information on add-on demand to target consumer-specific offers, but when consumers

are aware of their add-on demand,they now recognize a cheaper add-on. This allows uninformed

rivals to compete effectively: They can reduce the add-on price to attract add-on consumers

without adverse selection of base-good consumers. Awareness prevents adverse selection and

induces marginal-cost prices despite private information on add-on demand. I discuss below,

however, that the first main result also occurs in the setting without naifs when firms cannot

disclose add-on fees.

These results suggest that firms can use customer data to induce adverse selection of less-

profitable customers. This makes the data valuable even in highly competitive markets with ho-

mogeneous products.

2For example, naifs could underestimate their demand for an add-on service such as credit-card borrowing or late

payments, whereas sophisticates do not demand the add-on, that is, they do not borrow. I discuss further examples in

Section 7.

C

The RAND Corporation 2020.