Flawed math behind Burmese 'democracy'

What if you held an election and you weren't sure how many people showed up at the polls? To establish voting patterns and trends, one needs to have an accurate estimate of the population, clearly demarcated electorates and the eligible voters contained therein, and a system of tallying votes. It is not clear how closely these prerequisites were observed ahead of the November 7, 2010, vote in Burma* when population estimates in the country vary widely.

Burma has not had an effective nationwide census for decades. Previous ones took place during British colonial rule in 1931, under the post war social democratic government in 1953, and by the socialist government in 1983. The last official census in 1983 calculated the population to be 35.4 million, despite that count not being able to access considerable parts of the country due to civil war. But what is Burma's population now, and how did authorities count the votes on November 7, 2010? 

Population estimates between 2008 and 2010 vary from 44.2 million in 2009 according to the United Nations, based on Ministry of Home Affairs figures, to 59.1 million in 2010 according to the government's Ministry of Immigration and Population estimates. The figures stem from a survey of some kind conducted from 2007 in cooperation with the United Nations Population Fund (UNFPA) that estimated the population growth rate at 2.02% annually. From the lowest to highest figures over a two to three year period, there is a 15 million people differential in Burma government, UN and other international organizations' estimates. 

Other estimates are as widely divergent. The Rangoon-based UN agency, the Myanmar Information Management Unit (MIMU), released a map in 2009 with a breakdown of the population of all of Burma's 14 administrative units (states and divisions/regions) and the total population was 44.2 million based on ministry of home affairs figures. The 2009 Lonely Planet tourist guide claimed the population was 47.4 million. The US Central Intelligence Agency (CIA) estimates the population to be 48.1 million.

Many newspapers reporting on last year's elections in the past several months have variously estimated the population at 50, 51, 52, 54 or 57 million, all of the estimates likely based on Internet searches through the disparate figures on a variety of websites. According to data compiled by the United Nations Millennium Development Goals in 2008, the 2010 figure for Burma was projected at 50,495,000. In the 2008 Statistical Yearbook of Burma's Ministry of National Planning and Economic Development, the population figure was listed at 57,504,000.

So why the enormous fluctuation?

To understand the disparities in counting Burma's population, one has to know that estimating Burma's population is predominantly part of the system of authoritarian control. To monitor society, authorities have long employed a Draconian system of household registration.

Every house must have a list of inhabitants that are regularly reported to local authorities, at the suburb (or in Burma, the ward), or village level. Visitors are either denied permission to stay overnight or must be registered with the authorities. It is prohibited for foreigners to stay overnight in a private Burmese home and all hotel registration lists are reported daily to local police and immigration authorities.

This system of control is vertically integrated. Regular population numbers of small communities are relayed up to the next stage of monitoring control: from village to village tract, next to township, then to state or region level, and ultimately to national authorities. This is not unique to authoritarian systems but in practice it grants latitude to local authorities to act in any way to ensure that good news flows up.

In a country such as Burma, where avoiding the attention of authorities is a basic survival strategy, compliance by officials and citizens to accord with expectations is often the norm - despite questions of veracity or efficiency of the gathered information. Supplying positive news is an essential measure of loyalty in repressive states.

For example, the swift counting of the vote at the 2008 constitutional referendum - not an arduous task in a simple yes or no vote but still a challenge considering Burma's lack of development and infrastructure - was reached within a couple of weeks and publicly announced down to the individual vote: a 92% approval of the constitution with a 98% voter turnout. It was clear that local authorities knew they must deliver positive, even if false, news to the next layer of control all the way up to central authorities in Naypyidaw.

Most observers agree that Burma's electoral process in 2010 was one gigantic fix to ensure continued military dominance: Electoral laws were rigged and interpreted by a pro-military Electoral Commission; 2,200 political prisoners, including many opposition party leaders, were barred from contesting the polls; and the military's Union Solidarity and Development Party (USDP), with 18 million nominal members and a nationwide structure of offices and financial assets, ensured that smaller upstart parties were at a competitive disadvantage.

On voting day, November 7, widespread irregularities were reported, including the use of advance voting ballots to swing seats in favor of the USDP during the closing stages of counting ballots. The USDP won more than 77% of the seats in the two national level parliaments and a clear majority in the 14 regional and state based assemblies. The Electoral Commission announced that 22 million of 29 million eligible voters cast ballots, equating to a turn-out of around 75%-80%.

Inconvenient intervening variables
Few analyses of the Burma population include demographic factors that challenge the prevailing official assessment: displacement through conflict and development projects, migration to neighboring countries or urban areas inside the country, statelessness of the Rohingya Muslim minority, hinterland hill-tribes, and other marginalized populations, and haphazard or incomplete citizenship registration.

Burma is an extremely poor country sharing rugged and underdeveloped borderlands with Bangladesh, China, India, Laos and Thailand, simmering insurgent conflicts in its especially eastern borderlands, and a bewildering ethno-linguistic patchwork of peoples that official estimates put at 135 "national race groups". The following is a short list of inconvenient intervening variables that any accurate assessment of Burma's demographics must consider:

Internal displacement. For more than a decade, there has been a major problem of conflict and development induced displacement in eastern Burma. The annual survey of the Thailand-Burma Border Consortium (TBBC), an aid group that works with refugees, estimated 460,000 civilians were internally displaced in Burma in 2010 in a mixture of nominally government controlled areas, ceasefire militia enclaves and free-fire zones contested by state and anti-state forces.

These populations are a mix of recognized citizens of Burma and those whose births were not officially registered but have spent most of their lives under insurgent administration. Voting on November 7 was not held in several townships in border areas, such as the Wa special region in the north, and parts of the country where conflict is still raging in the Karen and Shan states.

Add to this tens of thousands of hill-tribe minorities, including Lahu, Akha, Palaung and others, especially in the northern states, who live on the fringes of state control and have never been officially counted. Ongoing conflict, or tensions between state forces and insurgents in border areas, meant that voting did not take place in parts of northern Shan State and eastern Burma, affecting parts of 32 townships (constituencies) and an estimated 350,000 people in total.

Refugees. There are more than 140,000 documented refugees in nine (unofficially) recognized camps in Thailand. These numbers have stayed largely constant since 1984 when the first major waves of refugees started to cross. More than 60,000 refugees have been resettled to third countries from these camps since 2005.

The ethnic Shan have only one very small recognized camp because most who have fled across the Thai border enter the migrant worker population. They easily number in the tens of thousands. India has approximately 50,000 ethnic Chin refugees in Mizoram and several thousand more in Delhi. Refugees also travel to Malaysia where some estimate 30,000-50,000 have fled to either work or apply for refugee status. Some refugees, however, retain their citizenship.

The Rohingya Muslim minority. Burma's most persecuted ethno-religious minority, the estimated one million Rohingya, have been the target of both large scale brutal military expulsions into Bangladesh (in 1978 and 1991) and denied citizenship and basic rights for three decades. Many Rohingya were paradoxically granted voting rights in 2008 and 2010 through the issuance of temporary ID cards and Rohingya political parties were permitted to contest the election (they were trounced).

Military-aligned Rohingya businessmen were permitted to contest and win seats through the USDP. An estimated 250,000 Rohingya live as refugees or as undocumented migrants in Bangladesh, while tens of thousands more work as migrant laborers in the Middle East and Pakistan.

Chinese migration to northern Myanmar. Chinese migration to Burma has demonstrably increased since the early 1990s, especially to Burma's second largest city, Mandalay.

Many Chinese migrants purchase citizenship, using business contacts with officials to secure it, while others are temporary laborers, such as the tens of thousands of road builders and dam construction workers now in Kachin State. There are no hard official figures on the size of this migration, but the presence of recent Chinese immigrants in northern Burma is clear to visitors and the source of periodic tension between ethnic Burmese and the new arrivals.

Migrant workers. Burma laborers leave their country in massive numbers, some for short-term work, others for many years. The standard estimate for Burma migrant workers in Thailand is two million. But in the absence of a fully functioning registration system, official Thai figures are much lower. In lesser numbers migrant workers from Burma also travel to Malaysia and Singapore, where working conditions are often marginally better.

Some of these workers were permitted to vote in last year's elections, if they were legally recognized as migrant workers, and cast advance ballots at Burma embassies. Some migrant workers refused to vote, fearing that officials would be able to extort money from them or their families back in Burma if they engaged with embassy officials.

Struck off household lists. Many people, especially Rohingya, who leave Burma because of persecution or for work are often struck off household registration lists because they have left the country illegally.

Many migrant workers leave their official ID cards inside Burma with their parents or family members because it is illegal to take the card outside of the country. Dissidents and others who have illegally left the country for clandestine training or work are regularly charged with breaches of the migration act and sentenced to long prison terms upon their return.

Internal labor migration. It became clear after 2008's Cyclone Nargis disaster that large numbers of landless laborers who had been working in the Irrawaddy Delta and may not have ever been counted either as temporary residents or as residents were amongst the official dead or missing count of 140,000.

The experience of the 2008 referendum, held soon after the disaster, is instructive in other ways. When Human Rights Watch interviewed survivors of the cyclone from 2008-2010, researchers encountered many who said they were not included in village household lists because of their isolated location.

Many said they were never given the opportunity to cast votes as local authorities completed their ballots for them. There are many other variables of transmigration not taken into account in official figures: how many people in Burma move within the country for work but fail to register with the authorities?

Authoritarian 'truth'. In his new book, Proofiness: The Dark Arts of Mathematical Deception, which details how governments and corporations bandy around deceitful figures, Charles Seife writes, "In skillful hands, phony data, bogus statistics, and bad mathematics can make the most fanciful idea, the most outrageous falsehood seem true. They can be used to bludgeon enemies, to destroy critics, and to squelch debate."

The obsession with numerical detail by Burma's authoritarian system is a prime example of what Seife calls "disestimation": granting credibility to a figure that is derived with too much uncertainty. Accurately estimating Burma's population is not only crucial for conducting elections - and one hopes that a genuinely free and fair election will take place one day in the country - but it is also crucial for increasing development projects and the disbursal of humanitarian assistance.

What Burma's new parliament, a reshuffled version of the former ruling military council, needs to give priority to in 2011 are credible population statistics that serve the needs of local development in health, education, land management, and economic reforms, including urgently needed micro-financing projects. These fundamentals have been lost in the haze of a system of control and the various responses by communities to survive under continued military rule (with a thin civilian facade for now).

If the United Nations system bandies around widely divergent figures, how will they coordinate with the national authorities and local communities to reach those most in need? Any agenda for international engagement with Burma should include reconciling the variables of communities that are not included on official registers and give more consideration to people who are used by the state when it suits them, and ignored when it doesn't.

David Scott Mathieson is a senior researcher in the Asia Division of Human Rights Watch.