Group identity in a network formation game with cost sharing

DOIhttp://doi.org/10.1111/jpet.12286
AuthorPritha Dev
Published date01 June 2018
Date01 June 2018
Received: 30 August 2016 Accepted: 9 November2016
DOI: 10.1111/jpet.12286
ARTICLE
Group identity in a network formation game
with cost sharing
Pritha Dev
IndianInstitute of Management, Ahmedabad
PrithaDev, Indian Institute of Management
Ahmedabad,IIMA Vastrapur,Ahmedabad, India
(pritha@iima.ac.in).
Iam extremely thankful to Jurjen Kamphorst for
hisinvaluable contributions to an earlier draft of
thepaper.
This paper introduces the choice of group identity, which is com-
prised of characteristics and commitments to these characteristics,
in a network formation model where links costs are shared. Players
want to link to the largest group given that linking costs for play-
ers of the same (different) characteristic are decreasing (increas-
ing) in their commitments. The equilibrium concepts used are Nash
equilibrium as well as one that looks for stability allowing for bilat-
eral negotiations. Conditions are shown under which the endoge-
nous choice of characteristics and commitments allow for multiple
groups with segregated or connected networks. When group iden-
tityis partly endogenous, such that characteristics are fixed and play-
ers only choose commitment, it is shown that the equilibrium popu-
lation commitment profile dictates whether the resulting network is
segregated or connected. When group identity is fully endogenous, it
is shown that multiple groups and segregated networks are possible
equilibrium outcomes but such outcomes are not stable unless the
group size additionally affects the costs of link formation.
1INTRODUCTION
This paper considers a model of network formation where the cost of a link depends on the group identity of players.
The study of group or social identity has been long incorporated into the social sciences through the pioneering works
of Mead (1934), Stryker (1968), Tajfeland Turner (1979), and Stryker and Burke (2000). Embedded in the literature
on social identity theory is the concept of identification with the social group as seen in Doosje, Ellemers, and Spears
(1995) and Terry and Hogg (1996). Identification measures the individual's feeling of connectedness with the group
where such identification might vary from being very strong to being very weak. With the insights from that litera-
ture in mind, in this paper the group identity of a player is captured by two variables; the first is a quantitative variable
capturing the group membership of the player, and the second is a qualitative variable called commitment capturing
the importance he attaches to that group membership. A person's group identity is thus the pair {group membership,
commitment}. Group identity is incorporated in a network formation game such that similar group memberships and
high commitments make the link cheaper to form, whereas different group memberships and high commitments make
the link more expensive to form. The network formation game analyzed is the standard connections model where
players want to link to the maximum number of other players at the minimum possible cost. The model allows for
390 c
2018 Wiley Periodicals,Inc. wileyonlinelibrary.com/journal/jpet Journal of Public Economic Theory.2018;20:390–415.
DEV 391
(a) Segregated (b) Connected
FIGURE 1 Differentcommitments, indicated by different shades, explain fragmentation in networks
(a) Segregated with Two Groups (b) Connected with One Group
FIGURE 2 Introductionof group identity, shape ={square, circle}, changes network structure
commitments and group memberships to be chosen simultaneously with the link offers. Hence, groups and the net-
work arise at the same time. The paper seeks to explore the conditions under which, instead of the entire population
choosing the same group membership and forming a connected network, multiple group memberships and segregated
networks are observed in equilibrium.
As an example,group memberships may involve players choosing political ideology: Republican or Democrat. At the
same time, players also choose their commitment to their chosen ideology.1Linkingto people with the same ideology
and higher commitments for the ideology is cheaper, whereas linking to a player with a different ideology and very
high commitment to the ideology is very expensive. If everyoneonly cared to belong to the largest possible network,
we should expect to see everyone choosing the same political ideology and there being one connected network. This
paper then shows, given ideologies, that the endogenous choice of commitment to ideology explains the division of
society by ideology. Further, this paper also shows, given that ideology and commitment to ideology are both choice
variables, it is possible for both ideologies to exist and for the network be divided along ideology if the costs of link
formation within a group vary by group size.
Figures 1 and 2 represent the main ideas discussed in the paper. The importance of the choice of commitment in
a network formation model is illustrated in Figure 1. The shapes Square or Circle represent the group membership,
whereas commitment is represented in the figure by different shades within a shape (darker shade =higher commit-
ment). The figure shows two panels with the same distribution of underlying group memberships but with different
1A Republican with very high commitment can be thought of as one who takesas the truth everything that is thought to be the Republican way of thinking.
On the other hand, a Republican with very low commitment believesin the basic Republican tenet of liberty, but might believe in other ideas not includedin
theRepublican way of thinking.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT