Census was ‘doomed’ from the start – and UN knew it

By Bill O’Toole

The Myanmar Times

April 7, 2014

Criticism of Myanmar’s census hit fever pitch last week when residents of Rakhine State were not allowed to self-identify according to their wishes, with even the United Nations appearing to turn on the government for its apparent back-flip.

But experts say the census was “doomed from the start”, and that donors and the UN had more than enough warning of the likely problems but did little to act on them. In particular, a risk assessment commissioned by donors “clearly warned” of many of the problems facing the program now, including flawed data and the inflaming of ethnic tensions, a person familiar with the report told The Myanmar Times.

The report was never released publicly and UNFPA did not respond to requests for information about its conclusions last week.

But there were many more public warnings about the census’ likely impact.

“Many individuals and organisations, both domestic and international, foresaw the obvious weaknesses and likely difficulties” of the census, said Fiona Dove, director of the Transnational Institute, a non-profit research organisation that released a report in February calling for changes to the process.

“From the beginning, there has been a lack of effective consultation and outreach … As a result, the census is not conflict-sensitive, and it is proceeding with flaws and deficiencies that should have been avoided.”

The chief complaint is on the question of ethnicity, and the problems are not confined to Rakhine State, where entire villages were passed over by enumerators because citizens wanted to self-identify as Rohingya, rather than the official term, Bengali.

Kachin civil society groups have been complaining for several weeks that the census questionnaire contained a list of sub-tribes that are unheard of in their state. Salai Lian Bawi Thang, the country program coordinator for the Chin Human Rights Group, said the census questionnaire contained numerous mistakes in the spelling of different Chin sub-tribes, of which there are 53.

“Indeed, the Chins widely do not accept that [there are] 53 Chin races/sub-tribes,” he said last week.

In response to events in Rakhine State, the United Nations Population Fund issued a statement saying it was “deeply concerned about this departure from international census standards, human rights principles and agreed procedures”.

“We are concerned that this could heighten tensions in Rakhine State, which has a history of communal violence, as well as undermining the credibility of census data collected,” it said on April 2.

However, observers, experts, and even donors to the census say the problems in Rakhine are indicative of broader flaws in the census methodology – flaws that the UNFPA and government were warned about early in the process.

In particular, many have criticised the UNFPA and the government for basing the census around the 135 officially recognised ethnic groups, which are an outdated legacy of the colonial period. Each of these groups was assigned a three-digit code, while respondents claiming any other groups would be counted as “other”, with code 914.

“The last British census in 1931 bequeathed an unreliable social and ethnographic map. But rather than addressing this unhelpful legacy, the present census is continuing many of these divisions and distinctions from the colonial and, subsequently, military government eras,” Ms Dove said.

A person familiar with the census preparations described the statements of concern from Western government as “disingenuous”, claiming that there was never a system in place to count those who chose to self-identify, in Rakhine or anywhere else.

“‘Rohingya’ data was never going to be collected in a way that could be reported upon. It was only going to be recorded as an ‘other’ foreign race, with the enumerator instructed to write the name, ‘Rohingya’, in the blank line next to the code box. There is no public record of if or how UNFPA [and Ministry of Immigration and Population] planned to tally handwritten responses to this question from the millions of completed census questionnaires,” the person told The Myanmar Times last week on condition of anonymity.

“The census was doomed from the start,” agreed David Mathieson, a Yangon-based researcher with Human Rights Watch. “[It was] predicated on a flawed ethnic classification, [included] overly cumbersome questions, and blithely ignorant of ethnic concerns throughout the country.”

Even several of the census’ donors, whose assistance likely made the process possible, said they had expressed concerns about the methodology to UNFPA in the months before the count.

Switzerland contributed US$3.2 million to the census, which was expected to cost around $74 million. A representative from the Swiss embassy in Yangon said the country was “one of the main promoters of UNFPA commissioning a risk analysis” of the census.

“Switzerland together with some other donors has consistently suggested that the census questionnaire be shortened and that the questions pertaining to religion and ethnicity be dropped,” said deputy head of mission and director of cooperation Peter Tschumi.

He cautioned, however, that it was “too early to speculate” on the “overall outcome” of the census.

“Once having the full picture of the result – that hopefully will be a positive one – we, together with other donor countries, will suggest further support measures to UNFPA and the government should the situation warrant it,” Mr Tschumi said.

The British embassy, which contributed about $16 million, stood by the work of the UNFPA and donors.

“Even with this serious disappointment [of data collection in Rakhine State] we judge that the census is still likely to be a more inclusive and valuable exercise than it would have been without international involvement,” a spokesperson said.

But Paul Cheung, co-chair of the census International Technical Advisory Board, which advised the census design and preparation, said it was “not uncommon” for governments to include ethnicity classifications.

In Myanmar’s case, the government insisted on designing the census around the 135 officially recognised ethnic groups, or “national races”, and agreed to add the “other” designation as a compromise with the UN and other donors.

“It is indeed quite common for the government to come out with a classification scheme. [There is] nothing wrong with that,” he said. “We allow countries to evolve their own arrangements and practices.”

The data from the census will not be released publicly until January 2015. Advocates say it is urgently needed to guide national planning, as a truly nationwide census has not been conducted since 1931. Estimates on the national population alone range from 48 million to 65 million.

Ms Dove from TNI said she agrees reliable data is important for a developing nation like Myanmar but said the current census will not provide it.

“Regrettably, criticisms have been ignored and difficulties have been treated as technical problems with simple, ‘one-size-fits-all’ solutions.”