Signaling about Norms: Socialization under Strategic Uncertainty

• Author: Fabrizio Adriani,Silvia Sonderegger
• DOI: http://doi.org/10.1111/sjoe.12240
• Date: 01 July 2018
• Published date: 01 July 2018

©The editors of The Scandinavian Journal of Economics 2017.

Scand. J. of Economics 120(3), 685–716, 2018

DOI: 10.1111/sjoe.12240

Signaling about Norms: Socialization under

Strategic Uncertainty*

Fabrizio Adriani

University of Leicester, Leicester LE1 7RH, UK

fa148@le.ac.uk

Silvia Sonderegger†

University of Nottingham, Nottingham NG7 2QX, UK

silvia.sonderegger@nottingham.ac.uk

Abstract

We consider a signaling model in which adults possess information about the dominant

social norm. Children want to conform to whatever norm is dominant but, lacking

accurate information, take the observed behavior of their parents as representative. We

show that this causes a signaling distortion in adult behavior, even in the absence of

conflicts of interest. Parents adopt attitudes that encourage their children to behave in

a socially safe way (i.e., the way that would be optimal under maximum uncertainty

about the prevailing social norm). We discuss applications to sexual attitudes, collective

reputation, and trust.

Keywords: Complementarities; conformity; cultural transmission; signaling; social norms

JEL classification:C72; D83; D80; Z13

I. Introduction

Family background affects disparate types of behavior ranging from

marriage choices and political attitudes to human capital investment

and civic engagement. The role of parental influence in determining

the behavior of children is the object of countless studies across

several disciplines. Within economics, a body of literature started by

the pioneering paper of Bisin and Verdier (2001) focuses on the

intergenerational transmission of culture.1The main innovation of this

*We thank two anonymous referees for their constructive comments, which have helped to

improve the paper. We are indebted to Chris Wallace for suggestions given at various stages

of the project, and to David Myatt for pointing out that a global games approach could be

fruitfully applied to our problem.

†Affiliated with the Centre for Decision Research and Experimental Economics (CeDEx).

1The body of literature on cultural evolution originally started outside economics, with the

papers of Cavalli-Sforza and Feldman (1981) and Boyd and Richerson (1985). Bisin and

Verdier (2010) provide a recent survey.

686 Signaling about norms: socialization under strategic uncertainty

body of literature is the reckoning that the socialization choices made by

parents are shaped by strategic motives. However, this body of literature

tends to emphasize the long-term dynamics of cultural evolution, leaving

the mechanism of transmission in the background. In this paper, we retain

the idea that parents are strategic but we take a different perspective. We

focus on a precise mechanism of parent–child socialization, which takes

the form of signaling. The underlying idea is simple: parents possess

superior information about the surrounding environment. Children thus

(consciously or unconsciously, through an as-if process) shape their beliefs

about society on the behavior of their parents.

Recent field and experimental evidence shows that parents distort their

behavior when observed by their children. Houser et al. (2016) find that

the tendency of parents to cheat in a coin-toss experiment is significantly

diminished when their child is in the room. Similarly, Ben-Ner et al.

(2015) report that parents increase their contributions in the dictator game

when these will be shown to their children.2This audience effect suggests

that signaling might play an important role in parent–child socialization.

The aim of our work is to understand how parent–child signaling affects

social norms. We consider a setting with a continuum of partially informed

players (adults) and uninformed players (children). Each adult has altruistic

motives toward his own child. Adults face a binary choice between two

variants of a cultural trait. After observing their parent’s action, each child

also faces a binary choice. An obvious example is one where children

must also choose which cultural trait to adopt, although, as we will see,

our set-up is more general and encompasses other applications.

An essential ingredient of our model is that the child’s utility from

either choice depends on the dominant social norm. This is, for instance,

the case when there are economic gains from adopting the prevalent

cultural variant, or in the presence of a conformist bias such as that

proposed by Boyd and Richerson (1985) in their theory of culture. For

example, honest behavior is typically more appealing when a social norm

for honesty is prevalent. Children form beliefs about the dominant social

norm by observing the behavior of their parents. The signaling game we

consider is thus anomalous, because senders (parents) convey information

about aggregate adult behavior in society (i.e., the dominant norm), rather

than signaling their own characteristics.

The second crucial ingredient in our story is the presence of strategic

uncertainty – that is, uncertainty concerning the behavior and beliefs

2The study also identifies what the authors call an emulation effect: after observing the choice

of their parent (or another adult), children modify their giving in the dictator game, in order

to narrow the difference. This suggests that children actively mold their behavior on that of

adults around them.

©The editors of The Scandinavian Journal of Economics 2017.

F. Adriani and S. Sonderegger 687

(and beliefs about the beliefs) of others. To model strategic uncertainty,

we borrow from the body of literature on macro-finance and adopt a

global games approach.3As argued by Morris and Shin (2003), “the

apparent indeterminacy of beliefs [that arises in many settings] can be

seen as the consequence of two modeling assumptions introduced to

simplify the theory. First, the economic fundamentals are assumed to

be common knowledge; and second, economic agents are assumed to be

certain about each other’s behavior in equilibrium.” Global games relax

these assumptions by modeling information in a more realistic way, thereby

escaping the indeterminacy straitjacket.

More concretely, in our set-up, aggregate behavior is not directly

observable. All adults receive private but correlated signals about the

underlying environment, and use their private information to predict the

information available to others (and, thus, their behavior). As in all global

games, each adult’s information can be very accurate, as long as it is

not perfect. As we will see, this grain of doubt over the behavior of

others allows one to pin down precisely which self-fulfilling beliefs will

prevail. In this respect, our model thus allows us to gain insight into a

notoriously elusive issue, namely the nature of the norm that emerges in

equilibrium.

We show that the desire to send the right signal to their child causes

adults to distort their behavior. We call this distortion the norm-signaling

bias. Intuitively, if the child knew the norm, then each adult could simply

take the action that maximizes his own expected utility. In contrast, with

signaling, adults have to take into account the fact that their behavior

might mislead their child into adopting the wrong behavior. Interestingly,

we show that this bias arises even when the interests of adults and children

are perfectly aligned.4

Our set-up allows us to characterize the signaling distortion in a precise

manner. To fix ideas, suppose that both parent and child must choose

between being honest or dishonest. If all parents are honest, a child’s

best reply is to be honest (e.g., because dishonesty would attract severe

sanctions or ostracism), while if all parents are dishonest, then dishonesty

is optimal for the child too. Consider now a society that is split in two

equal halves between honest and dishonest. What is the child’s optimal

behavior in that case? The answer to this question identifies what we

call the socially safe choice for the child. This is the action that would

3More precisely, our model draws on the body of literature on binary action global games

(Carlsson and van Damme, 1993; Morris and Shin, 1998).

4Our set-up thus dispenses with what Bisin and Verdier (2001) call imperfect empathy, namely

the idea that parents might want their children to adopt a certain cultural trait (i.e., their own)

even if this is not in the child’s interest.

©The editors of The Scandinavian Journal of Economics 2017.