The Myth of Power-Sharing and Polarisation: Evidence from Northern Ireland
| Published date | 01 February 2022 |
| Date | 01 February 2022 |
| DOI | http://doi.org/10.1177/0032321720948662 |
| Subject Matter | Articles |
https://doi.org/10.1177/0032321720948662
Political Studies
2022, Vol. 70(1) 81 –109
© The Author(s) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/0032321720948662
journals.sagepub.com/home/psx
The Myth of Power-Sharing
and Polarisation: Evidence
from Northern Ireland
Matthew Whiting1
and Stefan Bauchowitz2
Abstract
Whether power-sharing increases polarisation or not in post-conflict societies remains deeply
contested. Yet, we currently lack an adequate conceptualisation of polarisation to assess the
link (if any) between the two. This article offers a new conceptualisation of polarisation and
uses this to gather evidence from Northern Ireland to argue that the assumption that power-
sharing entrenches polarisation is not the reality that many think it is. By examining legislator
voting records, speeches by party leaders, manifestos and public opinion data, we disaggregate
polarisation into different issues, track it over time, and examine both elite and mass levels. We
find that overall polarisation declined, albeit some limited polarisation remained in cultural and
identity issues, but these were of low salience. We argue that this is the result of parties using
identity instrumentally for electoral distinction in a system of convergence – a process that is
independent of the effects of power-sharing.
Keywords
power-sharing, polarisation, Northern Ireland
Accepted: 17 July 2020
Introduction
Whether power-sharing increases polarisation or not in post-conflict societies is of central
importance to academics and mediators interested in managing and resolving conflict.
The rise in the number of negotiated ends to civil wars that rely on some form of power-
sharing (Billingsbo, 2013) makes this debate particularly relevant today. Yet, studies of
post-conflict societies lack an adequate conceptualisation of polarisation, which hinders
gathering satisfactory evidence to assess the link, if any, between power-sharing and
long-term polarisation. This article offers a new conceptualisation of polarisation in post-
conflict settings and applies this to gather evidence from the case of Northern Ireland to
1Dept of Political Science and International Relations, University of Birmingham, Birmingham, UK
2Governance and Conflict Division, GIZ, Bonn, Germany
Corresponding author:
Matthew Whiting, University of Birmingham, Birmingham B15 2TT, UK.
Email: M.Whiting.1@bham.ac.uk
948662PSX0010.1177/0032321720948662Political StudiesWhiting and Bauchowitz
research-article2020
Article
82 Political Studies 70(1)
argue that common assumptions that power-sharing entrenches polarisation is not the
reality that many think it is.
Debates around power-sharing have been unable to move beyond whether it is a valu-
able tool in conflict transformation or whether it merely institutionalises rival identities
that formed the basis of the conflict in the first place. For its supporters, power-sharing
neutralises conflict by guaranteeing the inclusion of minority groups while still adhering
to the principles of liberal democracy (McGarry and O’Leary, 2017). Power-sharing
reduces the likelihood of conflict recurrence (Hartzell and Hoddie, 2007); it incentivizes
violent spoilers to transform into stakeholders (Walter, 2002); it promotes parity of
esteem between competing groups (O’Leary, 1999); and, it encourages policy making
that meets the needs of as many in society as possible and not just the majority (Lijphart,
1969). What is more, when it comes to voting behaviour, it does not appear that power-
sharing inevitably leads to ethnic outbidding (Garry, 2016; Mitchell et al., 2009; Tilley
et al., 2019).
In stark contrast, critics argue that whatever gains power-sharing appears to offer,
these come at the long-term cost of embedding polarisation. From this perspective,
power-sharing pits fractious identities against each other and prevents cross-cutting poli-
tics from emerging that could dampen violent rivalries (Jarstad and Sisk, 2008; Roeder
and Rothchild, 2005; Tull and Mehler, 2005). By accommodating competing (usually
ethnic) identities, power-sharing actually incentivizes leaders to appeal to citizens in eth-
nic terms and strengthens the hand of hardliners (Horowitz, 1985). Power-sharing is seen
as creating a dynamic of outbidding by rivals in an attempt to gain electoral advantage,
pulling parties to extreme positions (Lee, 2009). This dynamic is compounded by the
nature of ‘involuntary coalition’, which insulates the executive from any effective opposi-
tion, thus creating a moral hazard for outbidding and polarisation (Taylor, 2006). Elite
behaviour is seen as having a spill-over effect on mass attitudes and behaviours, which
follow the lead set by elites (Zingher and Flynn, 2018).1
This long-standing debate has not been resolved due to the lack of an adequate con-
ceptualisation of polarisation in the post-conflict context. Existing understandings of
polarisation do not allow us to examine if some policy areas are more likely to be polar-
ised than others, how rates of polarisation may change over time, nor distinguish
between elite or mass polarisation. Instead, the term is often used in a catch-all way to
indicate a sense of deep political tension between rivals, but this inhibits careful empiri-
cal measurement. Therefore, this article offers a new conceptualisation of polarisation
and uses this to gather systematic evidence to answer the following questions: do divi-
sions between rival parties in a power-sharing system become entrenched over time or
even increase? Does polarisation emerge in some policy areas but not others? Does
designing a power-sharing system that accommodates particular identities prevent
broader politics from emerging?
To tackle these questions, we use the case of Northern Ireland, whose mixed record
under power-sharing makes it an ideal laboratory. This article challenges common
assumptions that power-sharing has increased rates of polarisation and, in fact, on most
measures it has declined. We find that polarisation is rare and when it happens it does not
impact the functioning of the institutions. Although some distance remains between the
parties in their electoral appeals around identity politics, this is not emphasised very much
and certainly less than bread-and-butter politics. Instead of polarisation, we suggest that
the parties are converging. There were also no indications of mass attitudinal or behav-
ioural polarisation. Occasionally, the parties emphasise identity politics for instrumental
To continue reading
Request your trial