Printer Friendly

Distorted, dangerous data? Lumyo in the 2014 Myanmar population and housing census.

Although Myanmar undertook its first census in thirty-one years in 2014, the governments of President Thein Sein and State Counsellor Aung San Suu Kyi have withheld the ethnicity statistics polled in the enumeration. The Ministry of Immigration and Population and the United Nations Population Fund framed the census question on ethnicity--or "lumyo"--to fail to produce reliable data, given that it neither met international standards nor took into account the complex and politically charged landscape of ethnicity in Myanmar. As a result, the ethnic results of the 2014 Population and Housing Census came to be viewed as destabilizing to the political reform and peace processes.

Keywords: census, ethnicity, lumyo, Myanmar 2014 Population and Housing Census, United Nations Population Fund (UNFPA), Myanmar.
The 2014 census is going to be the first true snapshot of the
population of Myanmar.

--Frederick Okwayo, Chief Technical Advisor on 2014 Census (2)

In 2014, Myanmar conducted its first census in thirty-one years, with the United Nations Population Fund (UNFPA) serving as the lead technical agency in support of what was then called the Union of Myanmar Ministry of Immigration and Population (MOIP). (3) Among the forty-one questions that the census asked was one on lumyo, or race, ethnicity or nationality. Although the data tallied on all other questions were released in a relatively timely fashion, both the President Thein Sein and the National League for Democracy (NLD) governments have withheld the statistics on lumyo. Why have these data been withheld from the public? In February 2016, the director of the census, a demographer who received his doctorate from Thailand's Mahidon University, gave one insight when he told the Myanmar Times that the release of lumyo data could "shatter the state's peace and stability" during the political transition (Nyi Nyi 2012, quoted in Pyae Thet Phyo 2016).

Why are lumyo data perceived to be so dangerous? When the numbers are released, many lumyo groups are going to be disappointed with their absolute and relative statistical representations, while viewing the statistics for other groups as suspiciously large, and perhaps artificially inflated by design. Because political representation and the fate of Myanmar's peace process is at stake, and because the data on lumyo also bear on issues of social status and deeply felt notions about ethno-racial hierarchies, all numbers will be suspect and probably criticized for distortion. This note argues that the UNFPA and the MOIP ignored the ongoing civil wars and constitutional provisions that politicized the collection of this information. They ultimately put at risk a fragile peace process and a tenuous start on what will be a very long democratic reform trajectory for Myanmar. They also selectively disregarded the very "global standard" on ethnic data collection to which both agencies committed themselves. Instead, the leadership of these agencies, supported by the foreign donors that funded the costly census, insisted on measuring identity in contextually inappropriate ways--in a country where people routinely risk their lives asserting or defending lumyo identity. In such a context, as the census director suggested in early 2016, hiding the results may have been seen as a "risk mitigation" strategy. He is likely to have recognized too late that the UNFPA/MOIP census methodology had led to a dangerously distorted "snapshot" of Myanmar's lumyo. (4)

Background to the 2014 Population and Housing Census

The impetus for taking a census at that particular moment in Myanmar's history is unclear. A census is not required under the terms of the 2008 Constitution. United Nations (UN) sources claim that Thein Sein requested UN assistance from Secretary General Ban Ki Moon on the sidelines of the 2011 East Asia Summit in Bali, (5) although advisors to the president remember UN staff promoting the idea before the Thein Sein administration had ever considered it. (6) However it came to be, a series of letters were exchanged between UN Special Advisor Vijay Nambiar and MOIP Minister U Khin Yi in April 2012. The letters committed the two parties to the use of the "global standards laid out in the Principles and Recommendations for Population and Housing Censuses, Revision 2, approved by the United Nations Statistical Commission" (Khin Yi 2012; Nambiar 2012). (7) In what was not an inevitable or necessary decision, Nambiar and Khin Yi landed upon these standards in particular among an array of possible choices of "international standards" by which Myanmar might be evaluated.

The United Nations and those funding the census justified the project as an urgent prerequisite to the country's policy reform process. The UNFPA and, later, donors of bilateral aid argued that it would be necessary in the development of evidence-driven policy. Nonetheless, the conducting of a census in Myanmar ran utterly counter to UNFPA's global guidance that censuses should be conducted only in countries at peace, (8) presumably not in a country that had suffered seventy years of civil warfare, still ongoing. Multiple categories of data collected--such as those on questions relating to immigration, literacy and ownership of taxable items--had the potential to trigger and exacerbate social tensions in such an environment. However, it was the lumyo question and the process by which the UNFPA and the MOIP administered it that became by far the most politicized and potentially destabilizing feature of this top-down process of capturing that "snapshot" of "Myanmar".

Foreign aid donors were initially cautious in reaction to the fundraising pitch that the UNFPA made in 2012. Those were the early uncertain days of the "transition", and there was also an awareness among donor agency staff that such a massive undertaking entailed risks that the UNFPA's and the MOIP's leadership were unable or unwilling to recognize. Eventually, however, as Western countries' exuberance concerning the peaceful nature of Thein Sein's economic and political reforms swelled, so did donor development budgets. Under end-of-fiscal year spending pressures, Australia abandoned what were considerable concerns about the proposed 2014 Population and Housing Census (PHC) and instead committed to funding it. The United Kingdom and several other bilateral donors followed, to augment government funding for a census that eventually cost US$75 million, with about US$15 million coming from the government of Myanmar. Not a penny of donor assistance went directly to that government. Rather, funding was channelled through the UNFPA.

Outside of the peace process, the census was the largest single non-infrastructure, donor-funded project of the Thein Sein era (2011-16). By way of contrast, the November 2015 general election received less than US$30 million in donor support.

The Structure of the Census

At the UNFPA's prompting, the census was planned to be the country's first-ever attempt at a de facto (9) hundred per cent headcount of every human being in Myanmar on the night of 29 March 2014. A force of about a hundred thousand teachers, as well as some midwives, would serve as "enumerators" (sayin-kauk). Enumerators attended training carried out by senior teachers and school principals--who served as local supervisors of the census. These sayin-kauk learned how to conduct the census interview and to fill out the Burmese-language forms. Trainers had in turn been taught by layers of others via a "cascade" training methodology, originally starting with a small number of trainers taught directly by the MOIP, the UNFPA and other agencies.

Township Immigration and National Registration Department (INRD) officials, who are law enforcement officers, were responsible for ensuring that the enumeration would be completed. Local census committees included law enforcement officers, welfare and administration officials, and untrained members of government-organized service providers, such as fire brigades and the Myanmar Maternal and Child Welfare Association. These local committees were responsible for planning enumeration, publicizing the census, numbering houses, rallying support, and--often--accompanying enumerators on household interviews to serve as interpreters and enforcers.

Teachers serving as sayin-kauk were given enumeration maps showing household locations produced by the MOIP with UNFPA oversight using tools such as Google Earth, but most ignored these maps as few knew how to read them (Stiegler 2014, p. 4). Each enumerator was to ask forty-one questions in 100 to 150 households over a ten-day period (Nyi Nyi 2012, slide 8). They were to repeat most of the questions over and over in order to obtain information on each member of a household. The enumerators were legally required to identify the head of the household, (10) usually the eldest male, as the "respondent". The work load could be daunting, with household size ranging from 2 to 4 up to as many as 80 people. (11) Local census committees and INRD law enforcement officers knocked on doors ahead of the enumeration period to warn residents not only to be home for the interviews but also to answer truthfully, as required by the July 2012 Population and Housing Census Law. Any answers providing "incorrect information" (hma-ywun-swa pye-khyin) was punishable by one month in prison and/or a fine of fifty thousand kyats (Ministry of Immigration and Population 2013, sees. 16 and 20).

Enumerators recorded most answers by filling in a bubble on one of the sixteen million large, machine-readable paper forms produced in 2013 by DRS, a British printing and data processing company that is likely to have absorbed much of the US$15 million the United Kingdom's Department for International Development contributed to financing the census (DRS 2014; Department for International Development Burma 2013).

What is little understood about the census interview or data collection process is that no respondent wrote her or his own answers on the census forms. Instead, the sayin-kauk asked the questions; consulted her handbooks, manuals and code lists; and wrote down the answers, sometimes directly on the form but many times in an unofficial notebook or stack of papers, from which she presumably transcribed the answers on to the questionnaire later. (12) Then, as per the field manual, the enumerator--not the respondent--was to check the responses for completion and accuracy, though, as one observation report noted, they "rarely" did so (Stiegler 2014, p. 14). Hence, no matter how a respondent verbally answered a question--on ethnicity, religion, occupation, type of cooking facilities--she or he had no knowledge of what the enumerator recorded.

Censuses in Comparative Perspective

Academics have long viewed censuses as deeply political undertakings, not simply "technical" procedures that produce apolitical, neutral, impartial data that inevitably lead to better policy. (13) The Nazi census, for example, was used specifically to build a case for the classification of populations for eventual extermination (Aly and Roth 2004). More recently, countries such as Canada, Macedonia, Albania, Afghanistan, Iraq and Nigeria have cancelled, postponed or annulled the exercise at various stages because of political disputes or questionable results, large-scale partisan and political contestation of data, and threats of or actual outbreaks of violence.

Census results do not produce neutral, unfiltered, true "snapshots" of any country. Instead, censuses distort, reinforce, legalize and freeze for posterity many arbitrary categories of class, citizenship and identity. The use of census data is as likely to have as many negative effects as it is positive for the very populations who are supposed to benefit from the "evidence" produced by a census. In the last thirty years, developing countries have found themselves pressured by donors to measure their populations every ten years according to international framing concepts, such as the categories and indicators of the Millennium Development Goals. The categories surveyed--usually defined in the English language and typically carried from one country to another with little contextualization or serious consideration of the politics of cultural and linguistic translation--serve to harden social complexities into manipulable social "facts" that are actually oversimplified fictions. If census technicians insist on describing results as a photograph of anything, it is one taken with a camera that focuses only on objects that are deemed by the wealthier nations of the world to be of significance.

The nature of such one-size-fits-all data collection means that census reporting relies heavily on large-scale metaphorical "photoshopping"--that is, reporting census outcome numbers and then immediately revising them on the basis of other more or less reliable data sources--in order to cover up the gap between the questionable census "results" and what ordinary people and elites understand as a more accurate depiction of reality. For example, on 17 January 2017, the UNFPA's Myanmar office issued a press release entitled, "Census Report: 4.25 Million People Born in Myanmar Now Live Abroad" (United Nations Population Fund, Myanmar 2017). In fact, the 2014 Population and Housing Census question on household members living abroad yielded a total of 2.02 million people, a number so obviously flawed that the UNFPA's own staff relied on the more systematic and contextually appropriate research of other international organizations. For some years it has been known widely that at least four to six million Myanmar-born people lived in other countries.

All forty-one of the questions on Myanmar's 2014 census produced problematic data, as international technical consultants, enumerators, tabulators and enforcers made thousands of day-to-day interpretive decisions that distorted allegedly neutral "snapshots" into social fictions. Most of this work involved no proper consultation with Myanmar social forces of any kind. Incomprehensibly, the UNFPA only brought Myanmar's robust and highly capable civil society groups into the process as a rubber stamp a few months before the enumeration went forward. (14) It took none of the groups' criticisms of the procedures to be followed on board to revise the census questionnaire or to adjust the major procedures.

The rest of this note focuses on just one census question category: "lumyo". Literally, this term means "kind of people or person", but in everyday usage it means "race". It considers this question category in the broader context of the many changes facing Myanmar when the census was planned and carried out.

The Broader Context of Recording Lumyo Identity

During the period of ongoing political transition since 2011, a combination of official and unofficial forces inside Myanmar, as well as activists in the diaspora and the international community, have been engaged in an unprecedented burst of ethnic and political boundary-making--narrating, counting, classifying, registering, documenting and identifying who belongs in the country and who does not, who is entitled to what, and who represents whom among the different nationalities and peoples. The category of identity known in Burmese as lumyo, which literally means "kind or race of people(s) or person(s)", is typically mistranslated as "ethnic" or "ethnicity" in English. In popular usage and in Burmese legalese, the concept is often preceded by the term, taingyinthar, an adjective that is close to the English concept of "indigenous". Historically, the concept of taingyinthar lumyo differentiates identity groups that most Myanmar people consider to legitimately "belong" both to and in the taing--the area long ago controlled by Burmese kings--of Myanmar and distinguishes taingyinthar people from those who do not belong (Transnational Institute 2014; Cheesman 2017). Underlying the concept of taingyinthar lumyo is a belief in the objective, biological, verifiable, fixed and blood-borne nature of identity. Lumyo is thus far closer to the English concept of "race" than to that of "ethnicity", as the latter connotes the possibility of some degree of social construction of identity labels. In describing those taingyinthar lumyo that foreigners typically call "ethnic minorities" (lunaysu) in Myanmar, most non-majority activists prefer the English terms "ethnic nationalities" or "national races". An open question is whether the Bamars, the ethnic majority, are one of the taingyinthar lumyo nationalities or national races. Among many Bamars, there is an assumption that those terms only refer to non-Bamars, as if Bamars somehow belong "more" or require less justification for citizenship identity claims, or as if they are so naturally of the taing as to require no designation.

Any census question about race asked during a period of civil wars fought in terms of identity, during a nascent and tenuous transition away from fifty years of authoritarian rule, and during a fragile peace process was bound to invite controversy and its results to be highly contentious. Additionally, the census occurred in the lead-up to a highly contested election in which dozens of political parties were organized around racial identities. However, despite advice to the contrary from experts on Myanmar, the UNFPA's technical advisors and donors refused adamantly to drop the question. They also rejected the use of the more accurate term, "race", or the more locally accepted term, "nationality", in any English-language discussions, publications, forms or settings. Against Myanmar connotations and meanings attached to the term taingyinthar lumyo the English translation of the enumerator manual for the 2014 Myanmar census specified, "Ethnicity is based on a shared understanding of history and/or territorial origins (regional and/or national) of an ethnic group or community as well as on particular cultural characteristics such as language and/or customs, beliefs or way of life" (Ministry of Immigration and Population n.d.c, p. 30). This definition had been copied verbatim from a 2006 UN Economic Commission for Europe (UNECE) census manual, which--unlike any UNFPA/MOIP instructions on Myanmar--at least conceded that "ethnicity" might be "sensitive". But ultimately the UNECE concluded that identity was measurable across cultures if addressed in technical, apolitical terms.

Measuring Lumyo Identity in the 2014 Population and Housing Census

By way of contrast to the approach to lumyo taken in the Myanmar census, the "international standards" handbook to which MOIP Minister U Khin Yi committed, the UN Statistics Division's Principles and Recommendations for Population and Housing Censuses (United Nations Statistics Division [UNSTAT] 2008, p. 140), reviews the complexity of ethnic, racial, tribal and other kinds of group identity. It concludes, "no internationally relevant criteria or classification can be recommended" (15) for the collection of group identity data. In other words, UN Special Advisor Vijay Nambiar and Minister U Khin Yi did not make binding promises to employ any international standard on the classification or criteria for recording self-identified lumyo: the set of standards to which they committed themselves explicitly rejects the possibility of a global standard.

The UNSTAT's Principles and Recommendations does recommend that, when questions regarding ethnicity are asked in a census, categories should not be pre-fixed. If, however, such categories are delineated in advance, each individual should have the option of choosing multiple self-identifications. The UNFPA technical advisors working with the Department of Population ignored this guidance from precisely the "global standard" allegedly adopted by the UN and the MOIP.

Thus, the forty-one-item census questionnaire, drafted in 2012 in English by UNFPA technical experts with no Myanmar expertise, was presented that same year in one consultation session with the government and one with UN agencies, but notably without popular representation in either session. It provided for an internationally replicable and locally distorted category of "ethnicity"--one completely in contradiction with the agreement signed between the government and the UN. Against their commitments to follow the standards set out in the Principles and Recommendations, the UNFPA viewed Myanmar's collection of data on "ethnicity" as a thoroughly technical matter and therefore one that was wholly unproblematic in political, social and cultural terms. This understanding was at odds with comparative experiences, even with one contemporaneous case in Afghanistan--a country suffering intense conflict along the very same lines of group identity. In that latter context, the UNFPA advocated the opposite protocol for its census questionnaire only one year before conducting the Myanmar 2014 PHC. In 2013, Laurent Zessler, head of UNFPA in Afghanistan, argued, "We don't ask for ethnicity or language spoken, this is on purpose. This country [Afghanistan] has so many issues to address between the political process, the economy and security, why complicate it?" (Quoted in Graham-Harrison 2013).

But in Myanmar neither the UNFPA nor the MOIP entertained any such context-sensitive critical thinking. For each individual, one three-digit code space was provided, with the expectation that census takers would refer to a code list of fewer than 999 ethnonyms. The census questionnaire form contained a blank line to the right side of the three-box lumyo code. Code 914 was "Other", designated for the sayin-kauk to write in by hand any ethnonym not found on what was then a yet-to-be-generated ethnic code list. Importantly, when the code list finally came out, a few months before the enumeration, "other" turned out to be a sub-category of "Foreign" races on the code list, as there was no space in the state or popular imagination for "others" who could possibly be taingyinthar.

The delays in producing the code list meant that no one knew how "ethnic" or nationality groups would be categorized or demarcated. When the list was finally produced, ethnic nationality activists and spokespeople found it problematic on numerous grounds. Census procedures called for information on respondents' lumyo based upon a much disputed list of "135" national races. That list of 135 is almost identical to one first deployed during the latter part of the Socialist era (1962-88) and resurrected during the early years of the military government (1988-2011) that succeeded it. As the Transnational Institute's February 2014 report on the PHC noted,
[I]n July 1989, the SLORC [State Law and Order Restoration Council,
the name of the junta at the time] chairman Gen. Saw Maung referred to
the "Census Department," from which he [said he] had discovered "135
categories" of "national race groups." At the time, with
constitutional government suspended, this was widely perceived as a
confusing but tactical attempt to weaken non-Bamar solidarity around
identity in a new game of "divide and rule". (Transnational Institute
2014, p. 6)

The list features the eight major race categories of Bamar, Shan, Kachin, Karen, Kayah, Mon, Rakhine and Mon, some of which it further randomly divides into subcategories on the basis of linguistic or cultural connections. But it also defines others by reference to the states named for large ethnic groups in which members of the smaller groups reside. For example, "Wa", who are in no way linguistically or culturally connected to Shans, are listed as a subcategory of "Shan" because of the population's location in Shan State.

The procedural decision on the part of the UNFPA and the MOIP that the 2014 census would record each individual as being of one and only one lumyo exacerbated the political ramifications of imposing the 135 suspect categories on the population of Myanmar. This decision did not align with the global standard set out in the UNSTAT (2008, p. 140) Principles and Recommendations, which holds that "respondents [should] have the option of indicating multiple ethnic affiliations". (16) Specifically, Principles and Recommendations notes, "Ethnicity is multidimensional and is more a process than a static concept, and so ethnic classification should be treated with movable boundaries" (United Nations Statistics Division 2008, p. 139). The collection of census data that allows for the representation of multiple identities is not only a common practice in other countries, but it is also in accordance with the UN standards to which the UN Secretary General's representative and the Minister of Immigration had committed themselves. Additionally, other documents issued by the Myanmar state--such as citizenship scrutiny cards and the household registry lists--historically had recognized multiple identities for the same individual. There is no record that the foreign technical experts working on the 2014 Myanmar PHC ever considered modifying the procedure for collecting data on lumyo to reflect either the international standard or the widespread experience of individuals with more than one identity. Hence, the 2014 PHC ruled out a "snapshot" that captured common mixed heritages such as Shan-Bamar, Mon-Karen, Sino-Bamar and others. Instead, when or if the lumyo totals are revealed, the data will present a highly racialized snapshot that distorts the depiction of every individual as having one--and only one--biological, cultural or lumyo identity.

Distorting Numbers: The Process of Enumeration

The "Ethnicity" question was number 8 on the census questionnaire, squeezed between questions on "Religion" and "Disability". Another controversial question, on citizenship status, followed it. The instruction to enumerators was to ask. the respondent, "What ethnic group in Myanmar does [NAME] belong to or for non-Myanmar citizens which country do you come from? [sic]", for each member of the respondent's household. (17) Only one code could be recorded for each person, and by law (Republic of the Union of Myanmar 2013, sec. 12[c]) the enumerators were to ask the head of the household--usually the oldest male--to declare an identity for all members enumerated. Any answers not on the list of 135, including foreign country designations, were coded as "Other", under the "Foreign" category. Enumerators were to hand-write in the "other" ethnonym.

The census took place in the politically charged lead-up to Myanmar's 2015 elections, in which taingyinthar lumyo data from the census should have been used for calculating which groups met the constitutional threshold for a population count that would give them minority representation in the state and regional hluttaw (parliaments). Aware of this provision, numerous ethnic activists attempted to rally their communities to answer with a singular lumyo code that would maximize their numbers. For example, Chin, Shan, Mon, and Zomi created lumyo-specific census committees to conduct surveys to pre-count their numbers, as protection from being undercounted by the census and thus losing out on potential political representation. (18) According to the head of the Mon census committee, "There is a Mon population of about 100,000 in Yangon [alone]", with many more living in Bago Region and Mon State, numbers that could have qualified Mon citizens for additional representation under the 2008 Constitution (Eleven Media Group, 9 February 2015). Other ethnic groups were concerned that the census was dividing them into smaller categories in order to take rights away from them. For example, some Kachin activists advocated for all Kachin sub-groups to identify not with sub-groups or clan names, but rather with the over-arching "Kachin" category, which was numbered 101 on the code list. They ran a campaign on social media and on signboards that read, "Kachin 101".

The field manual instructions were contradictory on this issue of respondents' indicating a major race or a smaller sub-group. On the one hand it insisted that enumerators were to record whatever ethnonym interviewees gave in response to the question. On the other hand it instructed enumerators to probe for subgroup identities if respondents named a major race, even if those major races were present on the code list.
For example if a person belongs to the Kachin main ethnic group and
Trone sub-ethnic group then according to the code list it will be
coded as "102". (19) If the person belongs to Paku within the larger
Kayin ethnic group, she/he will be coded "307". For "other
ethnicity/nationality" where the particular ethnic/nationality name
given is not given in the code list, you will code "Other"--"914" and
write the name of ethnicity or nationality in the space provided.
(Ministry of Immigration and Population n.d.c, p. 30)

This guidance was particularly confusing, since the "eight major races", as referred to in numerous legal and constitutional documents produced in post-independence Myanmar, had designated code numbers on the list of 135, but enumerators were told both (a) to accept whatever the interviewee said and (b) not to accept the larger "major race" category as an answer but rather to drill down to a subcategory. (20) A respondent could claim to be "Kachin", but be challenged by the enumerator to record a clan or sub-group name.

In late April 2014, this author conducted semi-structured interviews with nearly one hundred individuals who had been enumerated in the 2014 PHC. Interviewees were drawn from community groups in Yangon and Mandalay. The interviews focused only on how these individuals had responded to the census question about lumyo, although interviewees shared reflections on the census process more generally. Most interviewees reported that they had been asked only a small number of the forty-one questions on the census. A large percentage reported that, when asked "5a lumyo leh", they had answered "Myanmar". The latter was neither an ethnonym nor a lumyo category on the official code list. It is, rather, associated with citizenship. About a third of the interviewees who had answered, "Myanmar", said that enumerators had probed for a lumyo category from the official code list. Another third of the interviewees were asked, "So you mean you are Bamar?", and even among those who were not Bamar, they reported to me that they had said, "Yes, Sayama". And, perhaps more disturbingly, a third of those who responded "Myanmar" were not asked for any clarification by the enumerator, meaning that she made her own decision about what category to record. Enumerators had not asked more than thirty of those interviewed the question about their lumyo at all. Further, several self-identifying Muslims reported that they had answered, "Muslim", to the lumyo question, but, after I showed them that the code list of lumyo did not contain "Muslim", were surprised they had not been probed for a category. They did not know what enumerators had recorded as their lumyo.

And yet, at the end of each day, all enumerators were expected to return to the local census headquarters, review their forms, ensure that every question had a recorded answer, and turn their forms in to their supervisors. Hence, for all questions on the 2014 census, including that concerning lumyo, one should have serious questions about the reliability of the data. When it comes to the enumeration of lumyo, there are significant anti-democratic political implications of getting the data wrong.

The Political Implications for Minority Rights of Bad Data

To date, the government has announced several delays in the release, originally scheduled for 2015-16, of the data on ethnicity from the 2014 census. Before the census, officials warned that lumyo data would be reported as much as a year later than other statistics because of the necessity of "hand tallying" identities written into the space for the "Other" category. However, in early 2016 they released a far more complex set of write-in data in Volume 2-B of the 2014 Census, "The Union Report: Occupation and Industry" (Ministry of Immigration and Population 2016). This situation suggests that other motives have been at play. And, in fact, as early as September 2014, the rationale for the delay in releasing data on ethnicity shifted, when the UNFPA announced,
Regarding data collected on ethnicity there will be a consultation
process with different ethnic groups led by the government, after
which the results will be released to the public. These results are
expected to be released by the end of 2015 or in early 2016. (Ministry
of Immigration and Population n.d.b, p. 1) (21)

There is no public record of either the UNFPA or MOIP having undertaken such consultation, except for one vague news report stating that "The [Central Census] [C]ommission is in talks with ethnic groups who claim their numbers have been reduced in the census and ethnic names may be misspelled" (Eleven Media Group, 9 February 2015). (22) According to the then director general of the Ministry of Immigration and Population, U Myint Kyaing, who also served as a member of the Central Census Commission,
We could release [the data on ethnicity] only after a series of
discussions with ethnic representatives, the national races affairs
committee of Parliament, historians and population experts. We cannot
release it while the ethnic leaders have yet to confirm it. (23)

No process for any kind of confirmation had ever been laid out. But even if it had been, the rigid, single-identity categories used to collect lumyo data ruled out any significant opportunity for the public to challenge or reconfigure identity categories in politically or culturally sensible ways. No demographic photoshop tool exists that can make up for a politically naive and poorly designed question, one capable of producing only distorted data.

Promoted to permanent secretary of the Ministry of Labour, Immigration and Population in 2016 under the NLD government, U Myint Kyaing promised to release the lumyo data, but he has yet to do so (SAM Staff 2016). However, Commander-in-Chief Min Aung Hlaing, in his 15 October 2016 speech commemorating the one-year anniversary of the Nationwide Ceasefire Agreement between the Thein Sein government and eight ethnic armed groups, delivered very precise statistics--including figures that pointed to the potentially explosive consequences of the release of complete data on lumyo. (24) In it, he noted, that the majority of the population of two of the seven Myanmar states named for ethnic nationalities were not members of the lumyo whose names they bore. These were Kachin and Shan States.

The 2008 Constitution's delineation of the seven states named for ethnic groups is based on the assumption that the majority of their populations belonged to the ethnic group for which they were named decades ago. If Kachin State is not majority Kachin, however, the constitution provides for renaming and redelineating its territory. The constitution also establishes six additional "self-administered" units labelled with the names of single ethnic groups (Republic of the Union of Myanmar 2008, secs. 49-56 and 274-83). It has, however, no provisions for how to create future states or self-administered regions or zones named for individual lumyo or whether or how new units would be considered on the basis of ethnic population statistics--presumably from the census.

The 2008 Constitution does lay out a process for redrawing the administrative boundaries of states, regions and self-administered zones and divisions and for renaming them. The process is complicated and has not been tested, but as peace negotiations move generally in the direction of "federal" political reforms, it is likely that lumyo statistics--if ever released--will be used in claims for varying levels of administrative status and for territorial redelineation (ibid., secs. 53-55). Under the constitution, redemarcation requires that a majority of the voting-age population of a township to be administratively "moved" out of one state or region or out of one self-administered division or zone and into a new or different one approve the move. The "consent" of members of the state or regional parliament and then of seventy-five per cent of the national parliament must follow. Although these requirements involve significant barriers to redelineation, the mere prospect of it is exacerbating concerns about the forthcoming release of census data on lumyo. What if the enumeration procedures have produced statistics according to which Kachins, Mons or Kayahs are not in the majority in the states named for these groups? Perceptions of race, rights and territory may spur political movements for new administrative units named for various ethnic groups. Such a development would challenge the popular hegemonic post-independence narrative of "major" and "minor" races that underlies legal and informal regulatory, social and political relationships in Myanmar.

Additionally, the 2008 Constitution provides for minority "national race" representation on a population basis in the legislatures. Their participation is limited "mainly, to undertake their National races affairs" (ibid., sec. 17[c]). (25) Sections 161(b) and (c) of the constitution provide for guaranteed representation in all state and regional parliaments of lumyo minority ethnic groups in these territories whose population constitutes at least 0.1 per cent of the national populace. Both the Burmese and English wording of the provision is confusing. But, in the 2010 general election, the Union Election Commission (UEC) seems to have interpreted it to mean that "taingyinthar lumyo" populations inside a state's or a region's borders that amounted to about 57,000 people--plausible estimate of 0.1 per cent of the national population--should receive seats. It authorized twenty-nine constituency seats for "ethnic" or "taingyinthar lumyo" participation in state and regional legislative bodies. These included such previously undemarcated groups as Akha, Kayan and Lahu, as well as Bamar. (26) In 2015, when the 2014 PHC result had lowered the population count from an expected 60 million or more to 51.5 million, the UEC still delineated the same twenty-nine constituency seats. Again, no one knows where the data on lumyo populations used to make these decisions came from.

It is thus not known how the UEC calculated population totals to designate taingyinthyar lumyo constituencies. For the section 161 provisions in 2010, candidates for section 161 "ethnic" seats were listed on a separate ballot, and only those voters who had ensured that their names were on a lumyo voting list were allowed to vote for their own taingyinthar lumyo candidate on that ballot. (27) Thein Sein named lumyo representatives elected in 2010 "national races affairs ministers" in the state and regional parliaments. (28) That the same twenty-nine constituencies were replicated in the 2015 election suggests that the results of the 2014 census played no role in UEC designations. For example, the Mon Census Committee deployed volunteers throughout Yangon region in 2013 to conduct a survey that yielded a statistic of close to a hundred thousand Mons resident in the region (Lawi Weng 2012; Eleven Media Group, 9 February 2015). And yet, Mons did not receive a section 161 seat in that region, which has only two taingyinthar lumyo seats--for Rakhines and Kayins.

Hence, beyond the potential for the lumyo data to harm the peace process, there are further serious political implications of the insistence of a non-standard and deeply flawed attempt to collect this information. Both territorial and representation rights are at stake and should have been factored in when the question was vetted. Had popular input in the early stages of questionnaire drafting been sought, ethnic nationality civil society organizations and political parties would have flagged these issues for the UNFPA and the MOIP. If the resulting lumyo data is ever released, it is unlikely to square with anyone's expectations of the relative sizes of ethnic groups.


The attempt on the part of the Myanmar government, the UNFPA and foreign donors to collect data on ethnicity or lumyo was undermined from the start by choices that ensured difficulties in enumeration, flawed results and the potential for significant political fallout following the release of final data. By ignoring the highly political nature of any census, not to mention ongoing civil wars, and instead insisting upon framing the undertaking as a purely "technical" exercise, the UNFPA and the government created instruments for measuring and surveilling the population in the absence of significant consultation with the public or with stakeholders other than government and international organizations.

Any attempt to take a "snapshot" of Myanmar ethnicity or lumyo should have taken into account the complexity of hundreds of years of social history. The populations of many of Myanmar's villages, towns, districts, regions and states are ethnic mosaics, with varying degrees of diversity in social relations and settlement patterns. Over the past two decades, domestic and international migration, as well as civilian displacement and refugee flight, has surged, only adding to this fluidity as people have moved away at unprecedented rates from the traditions, languages and singularity of identity of family homes.

Finally, and perhaps most importantly, the UNFPA and the MOIP failed to adhere to the very set of standards to which they committed themselves. Had they adhered to these standards in the collection, analysis and dissemination of data on ethnicity or lumyo, a richly textured cultural "snapshot" of Myanmar may have been a boon to the country's reform process, particularly as political leaders and communities debate decentralization and federalism. As Principles and Recommendations (United Nations Statistics Division 2008, p. 139) lays out, "It is important that the responding public be informed of the potential uses and need for data pertaining to ethnicity". The UNFPA's and the MOlP's approach to public relations, however, entailed only top-down publicity campaigns, and hence served to disempower the very citizens who were promised the benefits of the census.

Mary P. Callahan (1)

Mary P. Callahan is Associate Professor, Jackson School of International Studies. University of Washington, Box 353650, Seattle, Washington 98195, USA; email:


(1.) The author was contracted to serve as a consultant to the UNFPA early in the planning process for the 2014 Myanmar census. She is the co-author with Daw Tin Tin Nyunt of "Political Risk Assessment: Myanmar 2014 Population and Housing Census", for the UNFPA, Yangon, 15 February 2013 (Callahan and Daw Tin Tin Win 2013); and with Christoph LeFranc and Nancy Stiegler, "Documentation of Risk Mitigation: 2013 Pilot Census, Myanmar", for the UNFPA, Yangon, May 2013 (Callahan, Stiegler, and LeFranc 2013). All of the research contained herein was acquired from public sources and post-contract interviews. A Myanmar-language version of this note was at time of editing due to be published in the May 2017 edition of Myanmar Quarterly.

(2.) Quoted in United Nations Population Fund, Myanmar (2013a).

(3.) Since 2016, it has been the Ministry of Labour, Immigration and Population.

(4.) This note focuses on the census process, procedures and data collection on the lumyo question in general. It does not document the series of decisions and choices made by the UNFPA, the MOIP and foreign aid donors that led to the withdrawal of census enumerators from Rohingya areas of Rakhine State, as that issue alone is worthy of an entire article.

(5.) In 2012, Ban Ki Moon (2012) reminisced: Last November, on the margins of the ASEAN-UN summit in Bali, President Thein Sein informed me that in 2014, Myanmar would conduct its first census in almost 30 years. He asked for United Nations' assistance in that effort, knowing of the great expertise of the UN Statistical Division and of the UN Population Fund, a longstanding partner. Notably, this request relating to the census was not mentioned at all in the actual record of this meeting (Ban Ki-Moon 2011).

(6.) Interviews with three former political and peace advisors to President Thein Sein, Yangon, 14 and 15 April 2014, and Naypyitaw 15 October 2015.

(7.) In committing to the Principles and Recommendations, Nambiar and Khin Yi designated a particular "global standard" rather than other possible sets of standards that may have been more appropriate to the context.

(8.) The UNFPA's main webpage on censuses (United Nations Population Fund n.d.) states clearly, "A census is among the most complex and massive peacetime exercises a nation undertakes" (emphasis added).

(9.) In the past, Myanmar's post-war censuses had always used the de jure method of conducting head counts, under which the legal residence rather than physical presence was the basis for the count.

(10.) Historically, men have been, both legally and in keeping with social norms, assumed to be the de facto heads of households (Gender Equality Network 2015, p. 75).

(11.) Interview, survey professional, Yangon, 20 December 2016.

(12.) The English-language version of the questionnaire is available at Ministry of Immigration and Population (n.d.a); for the Myanmar-language version, see <> (accessed 27 November 2014). It is not known how the unofficial records taken down by the sayin-kauk were processed--whether they were collected and stored by the MOIP and/or the UNFPA, destroyed, maintained by the INRD, passed on to Myanmar's General Administration Department, or held on to by the sayin-kauk themselves. The use of these unofficial documents for recording personal information represents a matter of serious concern about the confidentiality guaranteed in the census law and the global standards agreed to by the UN and the MOIP.

(13.) The classic formulation of how census categories politicize ethnicity is in Hirschman (1987).

(14.) It was not until 25 November 2013, four months before the PHC began, that civil society was consulted on the census. The UNFPA called a meeting in Naypyitaw, inviting members of civil society organizations, who listened to presentations and had the opportunity to air criticisms. It was not clear how these organizations were chosen. The UNFPA also set up a National Myanmar Census Advisory Committee (United Nations Population Fund, Myanmar 20136). It was only after the meeting, when the UNFPA posted this document, that attendees learned they had become members of the committee. As one attendee said, "Neither the invitation nor anyone at the meeting told us we were on any committee"; interview, Yangon, 5 December 2013.

(15.) Emphasis added.

(16.) Emphasis in the original.

(17.) Note, this question also defies UNSTAT Principles and Recommendations standards, which state explicitly, "Data on ethnicity should not be derived from information on country of citizenship or country of birth" (United Nations Statistics Division 2008, p. 139).

(18.) A year ahead of the census, Chin activists recognized the potential for under-representation in census statistics. Nge Pee, secretary of the Chin National Party, noted that out of the total of 135 ethnic groups, 53 were said to be "Chin". This deeply weakened Chin power, in his view. "Some can't even be accepted as ethnic groups as they hardly exist. We will compile an actual list to submit to the government", he declared (Eleven Media Group, 7 May 2013).

(19.) Not "101", which was the code for "Kachin".

(20.) Historically, the only major race not broken down into a variety of census categories has been "Mon".

(21.) The same document rationalizes the non-release of data on "occupation and industry" as being due to the necessity of "further manual coding".

(22.) The article implies that MOIP official U Myint Kyaing, head of the PHC, is the source, but it does not actually cite him as the source.

(23.) Quoted in United Nations Population Fund, Myanmar 2013b.

(24.) Min Aung Hlaing said, [M]ore than 1.6 million of population are living in Kachin State with 29.2 percent of Bamar, 23.6 percent of Shan, 18.97 percent of Jaingphaw, 7 percent of Lisu, 5.5 percent of Rawam, 3.33 percent of Lawwaw, 2.89 percent of Lacheik, 1.57 percent of Zaikwa, and 8 percent of other ethnics in population ratios in line with the census enumerated in 2014.... More than 5.8 million of population are scattering in Shan State with 35.23 percent of Shan, 11.44 percent of Bamar, 8.94 percent of Pa-O, 7.06 percent of Palaung, 6.41 percent of "Wa", 4.46 per cent of Danu, 3.22 percent of Kokang, 4.05 percent of Lahu, 2.35 percent of Jaingphaw and 16.84 percent of other ethnics. (Min Aung Hlaing 2016)

(25.) However, Section 262(i) suggests that they can also hold concurrently other ministerial portfolios in state or region governments and self-administered divisions and zones.

(26.) The full list: Bamar (5), Karen (5), Chin (3), Shan (3), Pa-O (2), Rakhine (2), Lisu (2), Akha, Intha, Kachin, Kayan, Lahu, Mon, Rawang (1). For locations, see TNI 2010, p. 8.

(27.) There is no record of how polling station officials dealt with voters with multiple lumyo identities on their Citizenship Scrutiny Cards (national identification cards). Additionally, some voters reported that they were denied the national race constituency ballot on the basis of arbitrary markers of identity (for example, whether one was wearing a longyi [sarong] associated with the national race on the ballot). Interviews, Taunggyi and Mandalay, November 2015.

(28.) Note that a similar guarantee of representation is offered to non-majority ethnic nationality communities within self-administered divisions and zones named for locally majority ethnic groups under Section 276(h) and (i). The threshold for such representation on the Leading Body of the division/zone is ten thousand.


Aly, Gotz, and Karl Heinz Roth. The Nazi Census: Identification and Control in the Third Reich, trans. Edwin Black. Philadelphia: Temple University Press, 2004.

Ban Ki-Moon. "Readout of the Secretary-General's Meeting with H.E. Mr. Thein Sein, President of the Republic of the Union of Myanmar". 19 November 2011 <> (accessed 1 May 2017).

--. "Secretary-General's Remarks at Census Event". 30 April 2012 <> (accessed 1 May 2017).

Callahan, Mary, and Daw Tin Tin Win. 2014 Myanmar Census: Political Risk Assessment. Yangon: United Nations Population Fund (UNFPA), 7 February 2013.

Callahan, Mary, Nancy Stiegier, and Christoph LeFranc. Documentation of Risk Mitigation: 2013 Pilot Census, Myanmar. Yangon: United Nations Population Fund (UNFPA), May 2013.

Cheesman, Nick. "How Myanmar's 'National Races' Trumped Citizenship". New Mandala, 1 May 2017 <> (accessed 1 May 2017).

Department for International Development Burma, United Kingdom. "UK Support for the 2014 Burma Census: Business Case--[pounds sterling]10 Million 2013/14 - 2016/17". July 2013 <> (accessed 17 January 2017).

DRS. "DRS Provides Technology for Myanmar's (Burma's) First Census in 30 years". 23 January 2014 <> (accessed 17 January 2017).

Eleven Media Group. "Chin Leaders Say They Will Compile an 'Actual' List of Chin Ethnic Groups Living in Their Respective State". Democracy for Burma, 7 May 2013 <> (accessed 19 May 2017).

--. "Ethnic Census Release Delayed". 9 February 2015 <> (accessed 19 May 2017).

Gender Equality Network. "Raising the Curtain: Cultural Norms, Social Practices, and Gender Equality in Myanmar". November 2015 <> (accessed 12 May 2017).

Graham-Harrison, Emma. "Afghan Census Dodges Questions of Ethnicity and Language". The Guardian, 3 January 2013 <> (accessed 19 May 2017).

Hirschman, Charles. "The Meaning and Measurement of Ethnicity in Malaysia: An Analysis of Census Classifications". Journal of Asian Studies 46, no. 3 (1987): 555-82.

Khin Yi. "Myanmar Census: Letter from His Excellency Minister U Khin Yi, Minister of Immigration and Population to His Excellency Mr. Vijay Nambiar, Under-Secretary General, Special Adviser to the Secretary General on Myanmar". 30 April 2012.

Lawi Weng. "Census Raises Issue of Ethnic Identities". The Irrawaddy, 31 August 2012 <> (accessed 19 May 2017).

Min Aung Hlaing. "Greetings of Commander-in-Chief of Defence Services Senior General Min Aung Hlaing at the First Anniversary of the NCA". Global New Light of Myanmar, 15 October 2016 <> (accessed 19 May 2017).

Ministry of Immigration and Population, The Republic of the Union of Myanmar. "2014 Population and Housing Census: Main Questionnaire". United Nations Statistical Commission (UNSD), n.d.a <> (accessed 19 May 2017).

--. "Myanmar Census Update September 2014: Release of Provisional Results". Myanmar Census 2014, n.d.b <> (accessed 17 January 2017).

--. "Population and Housing Census of Myanmar--Field Instructions Manual for Enumerators and Supervisors". Myanmar Census 2014. n.d.c <> (accessed 12 May 2017).

--. "Population and Housing Census Law". National Parliament Law 19, 29 July 2013.

--. "The 2014 Population and Housing Census--The Union Report: Occupation and Industry--Census Report Volume 2-B". March 2016 <> (accessed 1 May 2017).

Nambiar, Vijay. "Myanmar Census: Letter from His Excellency Mr. Vijay Nambiar, Under-Secretary General, Special Adviser to the Secretary General on Myanmar to His Excellency U Khin Yi, Minister of Immigration and Population". 30 April 2012.

Nyi Nyi. "2014 Population and Housing Census (Central Census Commission)". PowerPoint Presentation. Yangon: Myanmar Information Management Unit, 14 December 2012.

Pyae Thet Phyo. "Census Data Could 'Shatter' Transition Stability, Says Official". Myanmar Times, 24 February 2016 <> (accessed 19 May 2017).

Republic of the Union of Myanmar. Constitution of the Republic of the Union of Myanmar (2008). Yangon: Ministry of Information, Printing & Publishing Enterprise, 2008.

SAM Staff. "Myanmar Finalizes Ethnic Census Data without Rohingya". South Asian Monitor, 27 December 2016 <> (accessed 19 May 2017).

Stiegler, Nancy. "Union of Myanmar, 2014 Census Observation Mission: Final Report". Yangon, Myanmar: UNPFA, May 2014.

Transnational Institute. "Unlevel Playing Field: Burma's Election Landscape". TNI-BCN Burma Policy Briefing no. 3, October 2010 <> (accessed 1 May 2017).

--. "Ethnicity without Meaning, Data without Context: Identity and Citizenship in Myanmar". TNI-BCN Burma Policy Briefing no. 13, February 2014 <> (accessed 1 May 2017).

United Nations Economic Commission of Europe (UNECE). Conference of European Statisticians' Recommendations for the 2010 Censuses of Population and Housing. New York: United Nations, 2006.

United Nations Population Fund (UNFPA). "Census--Overview", n.d. <> (accessed 19 May 2017).

United Nations Population Fund, Myanmar. "UNFPA Addresses Preparations on Myanmar's Upcoming Census in 2014". 17 September 2013a <> (accessed 15 May 2017).

--. "UNFPA Set up National Myanmar Census Advisory Committee". 2 November 20136 <> (accessed 19 May 2017).

--. "Census Report: 4.25 Million People Born in Myanmar Now Live Abroad". 18 January 2017 <> (accessed 19 May 2017).

United Nations Statistic Division (UNSTAT). Principles and Recommendations for Population and Housing Censuses, Revision 2. Department of Economic and Social Affairs, Statistics Division. Statistical Papers, Series M, no. 67. New York: United Nations Publications, 2008 <> (accessed 19 May 2017.)
COPYRIGHT 2017 Institute of Southeast Asian Studies (ISEAS)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Notes & Comment
Author:Callahan, Mary P.
Publication:SOJOURN: Journal of Social Issues in Southeast Asia
Geographic Code:9MYAN
Date:Jul 1, 2017
Previous Article:Cities in Motion: Urban Life and Cosmopolitanism in Southeast Asia, 1920-1940.
Next Article:Moscow-Hanoi-Tirana relations in the context of the split in the "socialist camp".

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters