The contentious politics of the census in consociational democracies, ESRC

• Users may freely distribute the URL that is used to identify this publication. • Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. • User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) • Users may not further distribute the material nor use it for the purposes of commercial gain.


Impact Objectives
• Investigate the politics of the census in societies that are divided along ethnic, religious or linguistic lines • Extend existing knowledge of the relationship between the politics of the census and the design of political institutions in deeply divided societies • Offer insights into how policy makers can mitigate contentious political debates and polarisation resulting from the census

Dr Laurence Cooley, from the University of Birmingham, UK, is working on an ambitious project that seeks to uncover insights into the politics of conducting population censuses in post-conflict countries where the results have the potential to destabilise power-sharing arrangements
Please would you give us an introduction to the project?
The project investigates the relationship between the politics of population censuses and the design of political institutions in deeply divided societies, focusing on the case studies of Bosnia and Herzegovina, Kenya, Lebanon and Northern Ireland. In countries riven by ethnic, religious or linguistic divisions, power-sharing forms of government are often adopted following conflicts as part of peace agreements. This can involve the proportional allocation of power and resources to different groups based on their population shares, making the holding of a census highly political. I hope to uncover some important insights for future peace-brokers, policy makers, designers of censuses and international census observers to help navigate some of the many pitfalls associated with the running of censuses in divided societies.

How does your approach with this project differ from existing research in this area?
Most literature on the politics of the census focuses on the role of questions about race or ethnicity in shaping group identities and on the efforts of representatives of those groups to gain recognition through the census. A good example would be the emergence of a pan-ethnic 'Hispanic' identity in the United States, which researchers have shown was tied up with the introduction of the Hispanic category on the census. However, little has been written about census politics in deeply divided or post-conflict societies and the relationship between the census and the design of political institutions in these contexts is not well understood, so this is where the project will make an original contribution.
Please tell us about your background and how this led you to becoming involved in this project.
Previously, I had conducted research about power sharing in Bosnia and Herzegovina and it was the country's long-awaited first post-war population census in 2013 that first got me interested in the topic. Part of the reason why it took so long for Bosnia to hold its first census as an independent state was the anticipation that updated information on the relative share of Bosniaks, Serbs and Croats in the overall population might have implications for the way that the country's extensive system of power sharing operates. I had previously lived and studied in Northern Ireland and so was aware that censuses there have sometimes been controversial. The 1971 and 1981 counts were partially boycotted by some Republicans and more recent censuses have been accompanied by speculation by politicians and the media over the implications of the relative share of Catholics and Protestants in the population for the future constitutional status of Northern Ireland. Knowing a little about these two cases made me realise there was an opportunity for a comparative project that could contribute to understanding of the subject and highlight some elements of best practice.

What were the main challenges you have faced in this project? How have you overcome them?
The first challenge came when writing a 'pathways to impact' statement as part of the funding application, since my previous research was not designed to have a policy impact. Striking the right balance between demonstrating willingness and ability to engage with and possibly influence policy makers and having a realistic plan to do so was particularly challenging. In practice, this involved communicating with people in the policy world from the start of the project to ensure a degree of buy-in and maximise the potential of the research to generate findings that will be of use to them. The second challenge was trying to undertake extensive fieldwork in four countries in a relatively short time period. Despite the time pressures, it has been an instructive and rewarding experience.

POWER-SHARING AGREEMENTS
When a country's population is deeply divided on religious, ethnic or linguistic grounds, it is common for differences to be settled by the proportional allocation of government, civil and military positions to each of the opposing parties. Cooley says, 'Consociationalism is a particular form of power sharing that has been adopted through internationally-mediated settlements to internal conflicts. Often in these arrangements, each major group is The problem with such proportional allocation of power is that a new census may reveal changes in the make-up of the population, creating knock-on implications for the power-sharing formula. Therefore, when a census is mooted, disputes can arise between political elites and civic activists over the inclusion of questions about ethnicity, religion and language -or even whether these questions should be asked in the census at all. Once a census is designed, these same politicians and groups sometimes campaign to persuade citizens to answer questions in a particular way. Cooley adds, 'Although such powersharing settlements manage conflict by accommodating the interests of all groups, they can incentivise politicians to appeal only to their own group, which can lead to further polarisation. The census can get caught up in this process.'

THE NEED FOR A CENSUS
In each of the countries Cooley is studying, the census has been controversial, sometimes sparking intense political debate. In Lebanon, Cooley explains, 'There are fixed quotas of parliamentary seats for representatives from the country's main religious groups based on the population shares these groups held when the last census was conducted in 1932.' Everyone knows that much has changed since 1932, but there is too much at stake for group leaders to support another census that could have major implications for the quotas. In Bosnia and Herzegovina, meanwhile, there is a three-member presidency comprising a Bosniak, a Serb and a Croat. Cooley says: 'In the 2013 census, politicians were worried that the results would reveal that many people failed to identify with any of these groups, which could have called the arrangement into question.' In Northern Ireland, where politics is largely divided between those supporting Irish unification and those who wish to stay within the UK, these sentiments are often assumed to roughly align with religious denomination, and so the relative proportion of Protestants and Catholics measured by the census takes on highly political connotations.
Given these controversies, we could ask why censuses should be conducted at all? As a rule, censuses are conducted every 10 years and are designed on similar lines across countries. Cooley explains, 'Censuses are expensive exercises and a lot can change in a decade. There are other ways of counting heads, such as using administrative data already gathered by governments.' However, some post-conflict countries do not have the administrative capacity to use these alternatives and there are also cultural attachments to the census, as recently demonstrated in Canada, which tried to scrap it in favour of a household survey. A count of the population provides vital information for governments and is often the source of data necessary for holding elections. Therefore, there is international pressure on countries to conduct a census as soon as is practical after internal conflict.

INFORMING INTERNATIONAL POLICY MAKERS
Cooley's research has been conducted through extensive fieldwork, plus analysis of related media and official policy documents. He says: 'The research is qualitative and my main method of data-gathering is conducting interviews with key participants in the census process in each of the four countries.' The interview questions were designed following consultation with experienced colleagues, policy makers and international stakeholders in census-taking. He intends to maintain these links throughout the project and thus co-produce knowledge regarding the particular triggers of disputes over the census in consociational democracies. Fieldwork is now largely complete in Bosnia and Herzegovina, Northern Ireland and Lebanon, with a visit to Kenya currently being organised.

FUNDING
Funding support for this project was provided by the Economic and Social Research Council (ESRC). The work has already yielded interesting insights. Cooley states: 'Although the casestudy countries all have power-sharing arrangements in place, variation in the precise details of these arrangements can have a significant impact on the census.' He adds, 'Negotiators working to settle internal conflicts are concerned with stabilisation and managing internal divisions. However, it is often unclear how these arrangements can adapt to population changes over time.' Census results are influenced by migration and emigration, increasing levels of secularisation and the changing salience of ethnic and religious identities. Cooley's findings suggest that in some circumstances, campaigners try to actively encourage ethnic and religious identification through the census to prevent dilution of a particular group's share of power. Although civic groups often advise citizens of their right to self-identify when filling out their census forms, in particularly contentious cases and when there is a lack of trust in data privacy, scaremongering may influence people's responses to these questions.

CONTACT Dr Laurence Cooley
Cooley hopes his research will help inform policy around both the census and peacebrokering in post-conflict states. However, he says, 'I'm not under the illusion that this work will solve the problem of conducting censuses in divided societies. I am increasingly convinced that there is no such thing as an apolitical census, even in stable societies, since the very act of asking questions about people's religion, ethnicity, gender, sexuality or citizenship status is inherently political.' However, the project can highlight areas of best practice from comparative case studies and emphasise under-appreciated links between the census and power-sharing arrangements. Such information may help policy makers and civil society representatives more carefully consider the design and conduct of censuses in divided societies. Cooley adds: 'Given that censuses are often conducted with support from donor countries and overseen by international organisations, these insights will go beyond national boundaries and inform donor agencies and organisations engaging in developing statistical capability internationally, such as the UN.' l While censuses may appear to be just technical, statistical exercises, they are in fact inherently political. This is particularly true in divided societies, where group size may translate into political power