The importance of considering optimal government policy when social norms matter for the private provision of public goods

• Published date: 01 June 2020
• Author: Ingmar Schumacher,Guy Meunier
• DOI: http://doi.org/10.1111/jpet.12418
• Date: 01 June 2020

J Public Econ Theory. 2020;22:630–655.wileyonlinelibrary.com/journal/jpet630

|

© 2019 Wiley Periodicals, Inc.

Received: 21 August 2018

|

Accepted: 17 November 2019

DOI: 10.1111/jpet.12418

ORIGINAL ARTICLE

The importance of considering optimal

government policy when social norms matter

for the private provision of public goods

Guy Meunier

1,2

|

Ingmar Schumacher

3

1

Université, Paris Saclay, INRA, UR,

ALISS, Ivry‐Sur‐Seine, France

2

CREST, Ecole Polytechnique, Palaiseau

Cedex, France

3

IPAG Business School, Paris, France

Correspondence

Guy Meunier, INRA–UR1303 ALISS, 65

Boulevard de Brandebourg, 94200

Ivry‐Sur‐Seine, France.

Email: guy.meunier@inra.fr

Funding information

Chair Energy and Prosperity

Abstract

We study optimal government policy in a reference

model (Rege, 2004, Journal of Public Economic Theory,6,

65–77) of public good provision and social approval in a

dynamic setting. We show that even if complete

adherence to the social norm maximizes social welfare

it is by no means necessarily optimal to push society

toward it. We stress the different roles of social

externality and the public good problem. We discuss

the problem with the standard crowding in and out

argument and analyze the relationship with Pigouvian

taxes. We discuss the role of the cost of public funds and

show how it can create path dependency, the multi-

plicity of both optimal equilibria and optimal paths, and

discuss the role of parameter instability.

1

|

INTRODUCTION

The behavior of individuals is often determined by social norms,

1

which, among others,

2

may

help to overcome the underprovision of public goods. These norms can spread or vanish, and

often exhibit reinforcing mechanisms where the more people adhere to a norm the more likely

is its adoption for an individual. Hence policies that aim to correct for free‐riding and

1

Social norms concern a wide range of behaviors (from tipping to vengeance) and are not always welfare‐enhancing (e.g., dueling, foot‐binding). To contribute

to a public good is a special case of a “prosocial”social norm. Social norms have been shown to play a role for littering (Cialdini, Reno, & Kallgren, 1990;

Torgler, Frey, & Wilson, 2009), energy consumption (Allcott, 2011), recycling (Hage, 2009; Viscusi, Huber, & Bell, 2011), smoking (Nyborg & Rege, 2003b), fuel

economy (Yeomans & Herberich, 2014), or tax evasion (Frey & Torgler, 2007; Luttmer & Singhal, 2014). Farrow, Grolleau, and Ibanez (2017) provide an

overview of the theoretical approaches and empirical evidence related to proenvironmental behaviors. The dynamic interaction between norms and behaviors

has been recently empirically investigated by Huber, Viscusi, and Bell (2018) for recycling.

2

Other approaches have been considered, such as altruism (see Andreoni, 1990; Bergstrom, Blume, & Varian, 1986; Sugden, 1982; Warr, 1983), moral

constraints (Sugden, 1984), or descriptive norms and legal norms or rules (Bicchieri, 2006).

externalities should take into account social norms and their dynamic implications. While

intuition suggests that it is optimal to push society toward full adherence in case of a “good”

social norm, implying everyone becomes a contributor, this result has not been formally

established. The objective of the present article is to understand what is the optimal policy in

the presence of public good and a dynamically evolving social norm.

We follow much of the literature and consider that a “social norm”is a rule of behavior that

is enforced through social approval and disapproval (Elster, 1989). As a result of social

interactions, social norms can spread through society, and whether or not someone adheres to a

norm is the result of personal objectives and social ties. The crucial difference between social

norms and legal rules or moral norms is thus the social dimension which may give rise to a

potentially self‐enforcing dynamic of the norm. Such a mechanism can yield a multiplicity of

equilibria (Young, 2006, 2014) which translates into path dependency and historical lock‐in.

Concerning public goods, Rege (2004) stresses that a subsidy can help unlock society from a

zero‐contributor situation and push it toward a full contribution equilibrium. Nyborg, Howarth,

and Brekke (2006) provide a model of green consumption with social approval that shares

similar features. The existence of multiple equilibria associated with social approval has also

been studied by Lin and Yang (2006), who argue that only sizable subsidies may induce

significant shifts in the equilibria.

3

While these articles all discuss that both the individuals’decisions and the dynamics of the

social norm may usefully be directed through government policy, none of these articles study a

fully dynamic, optimal policy intervention. Rege (2004), for example, advocates for temporary

policies that help society above a tipping point after which the norm sufficiently penetrates

through society such that it develops its own positive dynamics. This view implicitly supposes

that there is a cost attached to this policy, as why would one want to stop an otherwise

beneficial policy. Thus we take the costly aspect of the policy intervention more serious and in

so doing we study whether or not the full contribution equilibrium really is the optimal

equilibrium. Based on the model in Rege (2004) we study the optimal government policy

intervention where the government maximizes a forward‐looking social welfare function taking

into account the preferences of the individuals and the dynamics of the social norm. We use

Rege’s model as, intuitively, it is easy to see why the full contribution equilibrium should be the

first best equilibrium. Second, it has a simple dynamic structure with a tipping point. This

tipping point is the crucial feature in the model without optimal policy and is still important if

one considers a simple subsidy, but in the model, with the optimal policy, the tipping point does

not play a role any longer.

We follow the literature on the theory of regulation (Laffont & Tirole, 1986, 1993) by

assuming a linear cost of public funds due to distortionary taxation (Ramsey, 1927). In addition

to the positive externality associated to the public good, “social”externalities, related to social

approval and disapproval, should be taken into account, and the dynamic of the diffusion of the

norm plays a crucial role. Our results show that with a linear cost of public funds, full

adherence is not optimal. The diffusion slows when the norm spreads, and it is not worth

further subsidizing to convert the few last noncontributors. We derive a Ramsey–Pigou formula

3

The models of Rege (2004), Nyborg et al. (2006), and Lin and Yang (2006) does not explain why there is a social norm to contribute but model its dynamics. The

approaches of Bicchieri (2006) and Young (2014), or Gordon (1989) and Bethencourt and Kunze(2019) on tax evasion, are similar. Others, within the immense

literature on the evolutionary roots of human altruism and prosocial behavior, have tried to model the emergence of social norms as evolutionary stable, fitness‐

enhancing strategies (e.g., Bicchieri, Duffy, & Tolle, 2004; Bowles & Gintis, 2009). Other dynamic approaches include, for example, status‐seeking

Dioikitopoulos, Turnovsky, and Wendner (2019).

MEUNIER AND SCHUMACHER

|

631