The 2017 by-elections, though limited in scope and political impact, were the first test of the NLD government’s commitment to its democratic transition. All actors, from the Union Election Commission to citizens, from political parties to civil society, took their responsibilities seriously.
Between January 26 and April 2, the People’s Alliance for Credible Elections (PACE) deployed more than 800 observers, data center operators and coordinators to monitor and assess several key aspects of the by-election process. PACE’s observers: conducted a survey of citizen perceptions of the voter list and voter registration process; monitored the public display and update of the voter list; assessed the campaign environment; observed the in-constituency advance voting process; conducted a Sample Based Observation of the election day process, from polling station setup to vote counting; and monitored the results tabulation process at the township level. PACE observers were present in all 22 townships participating in the by-election. Their professionalism was a testament to their commitment to Myanmar’s democracy and a key to the success of PACE’s observation.

Through its observation PACE was able to assess the quality and credibility of key aspects of the 2017 by-elections. Overall, PACE found the 2017 process to be credible, and that there was increased transparency in some components of the election process, notably the election calendar and the advance voting process. PACE also appreciates the UEC’s openness in sharing the list of polling stations, which was key for PACE to conduct a sample-based observation.

However, some of the recommendations shared by domestic and international observers following the 2015 national elections still remain relevant today. Some of these recommendations can be implemented directly by the UEC. Others require parliamentary changes to the current election framework. Along with other domestic observation groups, PACE stands committed to work with members of Parliament and the UEC to identify amendments that would make Myanmar’s elections more inclusive, transparent and accountable.

All Hluttaw elections are organized by the Union Election Commission (UEC) which is, under the Constitution, the only institution with the power to oversee political parties, to organize Hluttaw elections and to resolve electoral disputes. The UEC has the final decision on election-related matters. At the Union level, all members, including the Chair of the commission, are nominated by the President with the same five years term, and it is almost impossible for the Pyidaungsu Hluttaw to reject the nomination. Given the legal framework governing the appointment of the Union-level commissioners, there is no doubt that the transparency and inclusiveness of the commission remains in question. Moreover, the duties and responsibilities granted by the Constitution allows the commission to act unilaterally with no checks and balances.

Election District

Equal suffrage requires that every vote has the same weight irrespective of distinctions such as race, color, gender, religion, language, political opinion, property, birth or others. Under the Constitution, the election districts of the Amyotha Hluttaw are based on states and regions, regardless of size and population, and for the state/region Hluttaws, townships are divided into two districts equally. For Pyithu Hluttaw elections, the Constitution allows the UEC to draw the boundaries based either on population size or simply on the current township administrative boundary. Currently, the Pyithu Hluttaw’s election constituencies are based on the township administrative boundary, resulting in significant differences in the value of the vote among constituencies. In the 2017 by-elections, the size of Pyithu Hluttaw constituencies ranged from less than 45,000 eligible citizens to almost 380,000, making some votes weigh more than eight times more than others.

Voter List and Voter Registration Process

A complete and correct voter list is fundamental to ensure that all citizens are able to exercise their political rights during elections. An inclusive voter list and a transparent registration process promote public confidence in the electoral process. The current list is the result of the UEC’s efforts prior to the 2015 elections to compile up-to-date information from the General Administration Department (GAD) and Department of Immigration (DoI), as well as a door to door update process that was supposed to reach all households in the by-election areas prior to the 2017 elections. The resulting list was then displayed to the public to identify and request any needed changes. However, the door-to-door update reached a far smaller percentage of households than expected, and the engagement of citizens in the public display process was lower than desired. Given this experience, and the extraordinary cost of updating the list with GAD and Department of Immigration’s information, the quality of the list and the lack of transparency of the voter list is a source of concern for the next general elections in 2020.

Voter Education

Without proper voter education, elections cannot be considered inclusive and transparent. Voters have the right to access all information related to the candidates and the electoral process. Even though the UEC has the responsibility of announcing the timeline of all election activities, and both the UEC and civic groups conducted voter education both in 2015 and 2017, there is no legal provision for the UEC to conduct voter education. Moreover, while it is worth to acknowledge the UEC’s efforts to use some minority ethnic languages in voter education materials ahead of the 2017 by-elections, there are still no specific guidelines to promote minority ethnic languages in voter education nationwide.

Campaign Resources 

Currently, the Hluttaw election laws do not establish limits on individuals’ contributions to political campaigns. This could result in a campaign being funded primarily, or even only, by a single donor, creating a conflict of interest and threatening the independence of elected representatives. The laws also establish a single cap for campaign expenditures for candidates for elections at all levels - Amyotha, Pyithu and state/regional Hluttaws -- regardless of the size of their constituencies. This puts candidates contesting in election districts with a bigger population at a disadvantage.

While it is important that electoral reform results in a legal framework that meets the principles of democratic elections, it also is important that the reform process itself should be inclusive, transparent and accountable. The amendments to the laws should not be agreed only by members of the majority party in Parliament, but should take into consideration the views of all political parties and the voters.

The political campaigns are the most important aspect of the pre-election period. Whether candidates were treated equally under the laws, had equal access to public spaces, were allowed to express their message freely, and were allowed to reach out to their constituents impact the fairness and credibility of the elections.

From February 1 to March 30, PACE deployed long-term-observers (LTOs) to all 22 townships to assess the transparency, accountability and inclusiveness of the campaign environment. Even though there were isolated incidents reported in the media, in general the campaign environment was calm and orderly. Candidates were able to conduct their activities without any major incidents of intimidation or violence. All campaign requests were approved by the election authorities without any change to time or location. There were no reports of interference during the campaign, or of incitements or comments regarding candidates’ gender, religion or race. However, compared to the 2015 elections, candidates’ campaigns were much less active during this by-election process.

The following detailed findings represent the viewpoints of citizens and individual candidates as expressed in 1,464 interviews conducted by LTOs and direct observation by PACE LTOs at 338 rallies. This information does not include activities or viewpoints of party headquarters, other party supporters or other groups. It also does not include information about activities conducted by parties or candidates before the official campaign period began. The findings are also based on 173 interviews the LTOs conducted with members of the local election sub-commissions.

Campaign Activities 


  • The most common outreach activities reported by the candidates PACE interviewed were distributing materials (48%), holding rallies (38%) and hanging posters (34%). The next most common forms of outreach were door-to-door outreach (26%) and parades/loudspeakers (20%). Very few candidates said they used technology or media to reach voters; only 3.5 % said they used media (interviews) to reach voters. and 6% using social media/Facebook. Almost one quarter of candidates (23%) said they did not have any activities planned. This data does not capture outreach activities conducted by parties’ central committees or by other party supporters. 


  • Among those interviewed, candidates from the USDP and NLD were the most active conducting campaign activities. Parades and loudspeakers were most used by NLD (46%), followed by other big parties (30%). A significant number of candidates from small parties (26%) and other big parties (19%) responded that they did not plan to conduct any activities at all.

Campaign Rallies


  • Of those interviewed, candidates did not report many problems in the rally approval process. All candidates from four party categories said their rallies were approved without changes. Two out of five candidates (40%) said their campaign requests were approved one day before or one on the same day the campaign activities were planned. 


  • Of rallies observed by PACE, most were held at private offices/homes (48%), party offices (21%), religious places (10%), and public spaces, like markets or parks (8%). Very few campaign events were held in sports stadiums/fields (2%) or government buildings (3%). None of the observed rallies were held at industrial places.
  • Candidates from all party categories were more likely to use private offices or houses to conduct campaign activities compared to other places. On the other hand, big parties other than the NLD and USDP, and small parties and independent candidates were more likely to use public spaces (like parks, markets, etc.) than the USDP or the NLD.
  • At rallies observed by PACE, big parties (15%) (other than the NLD and USDP), were more likely to hold rallies in religious places.
  • At most rallies observed, candidates handed out printed materials (57%). Other goods, like food (14%), party souvenirs (22%) were also handed out. Nothing was provided at 21% of rallies observed. Small parties and independent candidates (73%) were more likely to hand out printed materials compared to candidates from the other categories.

Campaign Messaging


  • Party leaders joined as speakers at 44% of the observed rallies, celebrities at 5% and community leaders at 3%. At more than half (51%) of the observed rallies there were no speakers besides the candidates.


  • At observed rallies, candidates from USDP, NLD and other big parties had similar rates of party leaders join the rallies. Independent candidates and those representing small parties were more likely to campaign without another speaker.
  • At all of the observed rallies, no speaker made any comment about a group or person based on their religion, race or gender.

Interference in Campaigns

In candidate interviews, PACE asked candidates if they faced any problems in the campaign (like interference in campaign activities, problems with job/business, physical threats/harm, property/campaign material damage, problems with friends/family or bribes, etc.)

  • Nearly all (96%) candidates said they did not face problems.
  • PACE did not witness interference or disruptions at any of the observed rallies, regardless of what kind of party was conducting it.

Voter’s Perception

    • Of the voters interviewed, 62% said voters around in their areas were aware of the by-elections, and 57% were interested in voting in the by-elections.


  • A majority (86%) of the interviewed voters said they could vote for any candidate they wanted with any fear of repercussions.



PACE deployed observers to 22 randomly selected wards and villages across the by-election areas during eight of the 14 days of the public display of the voter list (February 1 and 14). During its observation of the public display of the voter list, PACE witnessed that a small number of voters visited the centers to check their information and that very few people submitted forms to request changes.

PACE recognizes the Union Election Commission’s (UEC) efforts to ensure that all display centers were open to observers and all materials necessary to make changes in voter list were available to citizens.

Voter Education

PACE observers reported voter education materials at most locations around the display centers, but did not report seeing any voter education training or meetings. As observers only monitored within the center and its immediate vicinity, it is possible that voter education training or meetings could have been conducted in other locations without the observers’ knowledge.

  • There were voter education materials at most locations (82%).
  • PACE observers saw posters in 76% of locations and pamphlets in 25%. In 11% of locations, PACE observed the use of loudspeakers. PACE observers did not witness any voter education meetings or training during the period of observation.
  • In more than half of the locations (52%), no actor was observed conducting voter education activities. In one third of the places (34%), the local election sub-commissions were observed conducting voter education activities; in 18% of the places, CSOs were observed conducting voter education.

Display Center Management and Materials

To understand the administrative procedures of the centers, PACE observed their layout, the presence of materials, opening hours, and the behavior of display officials to ensure that changes requested by voters would be processed according to the UEC’s guidelines.

  • Most centers (95%) started the voter list display on February 1 as scheduled and opened the updating process to the public all the time during the observation period.
  • PACE was allowed to observe in all centers.
  • Most centers (94%) had all necessary forms and displayed the voters list.
  • In 81% of the centers, lists were displayed so that all voters, including elderly and disabled, could easily see the list.
  • At 94% of the centers, display officials were present. Of the officials present, 93% provided assistance to people who required it.

Presence of Political Parties and CSOs

During the first week of the display, PACE found that at most centers there were no political party representatives or other CSO volunteers present.

  • At approximately 86% of the centers, PACE did not see any political party representatives. The NLD was present at 9% of the centers and the USDP at 7%.
  • At approximately 80% of the centers, PACE did not see any other civil society volunteers.

Turnout and Submission of Changes

In the centers where PACE observed, volunteers saw quite modest turnout overall.

  • In 91% of centers, PACE observed that only a small number of people (0-50) came to check their names during the observation.
  • In 57% of centers, PACE did not observe any voters submitting forms. PACE cannot say why voters did not submit forms. In 35% of centers, only a few people (1 to10) submitted forms to make changes to the list.

Intimidation and Interference

A safe environment is one of the most important factors contributing to voter turnout. PACE observed whether any intimidation occurred in and around the centers. In nearly all the centers, PACE observers did not witness intimidation of voters or interference by unauthorized persons.

  • In nearly all centers, PACE did not see any intimidation of voters.
  • In nearly all centers, PACE did not see any interference by unauthorized persons in the process.

In-constituency advance voting provides the opportunity to vote to citizens who are not able to go to their assigned polling station on election day. It is important that the process is transparent and inclusive, and that citizens can cast their ballots in secrecy, and with no interference or intimidation. On the other hand, it is also important to make sure that the procedures are designed and implemented in a way that minimizes the risk of fraud.

In the days prior to election day, PACE deployed long-term observers to 22 randomly-selected wards and village tracts to monitor the in-constituency advance voting process. PACE assessed the quality of both the mobile voting and the process conducted at the sub-commission office. Generally, PACE found the process to be more open and transparent than it was during the 2015 general elections. The UEC announced the exact schedule of the voting, and both domestic and international observers were able to monitor the process without any restrictions. The secrecy of the vote was respected both at the sub-commission office and in mobile voting. However, there were isolated cases where citizens casting advance votes were not recorded in the form used by polling station officials to determine whether citizens had already voted, slightly increasing the risk of double voting.

On election day, PACE deployed 595 short-term observers (STOs) to 22 townships in eight states and regions to observe the election-day process, and an additional 22 long-term observers (LTOs) to observe the results tabulation process at township tabulation centers. Overall, the whole election day was calm, orderly and well administered.

The opening and setup process was transparent, and almost all polling stations followed the UEC’s procedures, regulations and instructions. At very few polling stations, some important materials were missing and very few people were voting on behalf of others. Almost all polling stations started voting by 6:30 am and very few polling station started voting after that time.

The voting process was transparent to citizen observers and party agents, and nearly all voters were able to cast their ballots without any interference or intimidation. However, PACE found that there were isolated incidents of citizens who were unable to cast their votes because their names were not on the voter list, reducing the inclusiveness of the elections.

PACE noted that the closing and counting process was more transparent to observers and party agents than during the 2015 elections. In the by-elections, more polling stations publicly posted the results forms, and the counting procedures and the decision criteria for invalid votes were consistent and accountable.

Polling station setup 

To assess the quality of the full process, it is important that observers are allowed to be present at the polling station before voting and observe as polling station officials set up and prepare for the voting process. Accordingly, PACE deployed its short-term observers to the polling stations before the opening and instructed them to observe the preparation process. However, the polling station official guidelines did not contain any specific instructions to open the setup and preparation process to observers or other members of the public. Instead, the guidelines require that the empty ballot boxes must be shown to “witnesses” before the voting begins. Even without clear guidelines from the UEC, PACE was able to observe the setup and preparation process thanks to the openness and transparency of polling station officials. Generally, the setup and preparation process was transparent and nearly all polling stations were accessible to elderly and disabled voters.

Detailed Findings

  • 99% of observers were permitted to enter the polling station before voting began. Compared to the 2015 general elections, 2017 by-elections were more transparent (94% of observers were allowed to observe the opening in 2015 general elections). The isolated incidents of observers who initially were not allowed to monitor the setup procedures were quickly addressed with the assistance of election authorities.
  • 88% of polling station facilities were accessible to all voters, including elderly and disabled voters.
  • At least 10 polling station members were present in 88% of polling stations.
  • In 93% of polling stations, the advanced ballot boxes were delivered before the station opened.
  • In 85% of polling stations, Form 13 (Advance Voting) was posted outside the polling station.
  • At the time of opening, 96% of polling stations had all necessary materials, while 3% were missing results forms 16 and 17, and 1% were missing voter lists. Less than 1% of polling stations were missing indelible ink.
  • In 92% of polling stations, voting began by 6:30, while in 8% voting began after that time.


Voting Process

Observation of the voting process can tell if the election day process has met democratic standards and principles in terms of transparency, accountability and inclusiveness. While the presence of observers and party agents demonstrates the level of transparency, the presence of unauthorized persons can undermine the level of the accountability of the polling station officials. It is important that all eligible citizens are able to exercise their political rights on election day. Therefore, it is important for the polling station officials to make sure that all voters whose names were on the voter list be able to cast their votes secretly and freely, and that people whose names were not on the list not be allowed to vote. The level of accountability of the polling station officials on election day can be assessed by observing if they followed the laws, by-laws, regulations and guidelines in a consistent manner. Overall, the voting process was transparent, accountable and inclusive. However, PACE observed that a few people were turned away because their names were not on the voter list.

Detailed Findings

  • Party or candidate agents were present during the voting process in 86% of polling stations across the by-election areas, but coverage was higher in Yangon (94%) than in other locations (79%). Agents for the USDP were present in 79% of polling stations and agents from NLD in 77%, while agents from other Burman parties were present in 8% and ethnic parties in 19%. Agents from independent candidates were present in 2% of polling stations. Representatives from NLD and USDP had greater presence in Yangon than in other locations, while agents for ethnic parties had more presence outside of Yangon.
  • In 91% of polling stations, there were no unauthorized people present. However, in 9% of polling stations, unauthorized persons were present.
  • Voters were asked to show proof of identity documents (such as a voter slip or National Registration Card) at 92% of polling stations.
  • In 25% of polling stations, less than 10 people who came to vote were turned away because they were not on the voter list. This number was higher in Yangon (35%) than in other locations (16%). In 2% of stations, between 11 and 50 voters were turned away because they were not on the voter list.
  • In 94% of polling stations, every voter whose name was on the voter list was allowed to vote. However, in 5% of stations, a few voters whose names were on the voter list were not allowed to vote.
  • In 97% of stations, only people with names on the voter list were allowed to vote. However, in 3% of stations, people whose names were not on the voter list were allowed to vote.
  • Voters were able to cast their votes in secret in 98% of polling stations.
  • In 98% of polling stations, there was no intimidation or harassment of voters inside or in the immediate vicinity of the polling station.
  • In almost all stations, voters’ fingers were marked with ink as they left the premises.
  • Police were present outside 80% of polling stations.
  • Observers were allowed to fully observe the voting process at 97% of polling stations. At fewer than 4% of polling stations, observers were allowed to monitor the process, but with some restrictions.

Closing and Counting

In order to have a comprehensive assessment of the whole election day process, observers must witness the closing and counting process once the voting process has been completed and the polling station has closed.  The level of transparency can be measured by how open the closing and counting process is to citizens, party agents and observers. On the other hand, the fact that necessary procedures, such as determining whether ballots are valid or invalid, are taken consistently during the counting process determines the level of accountability. On election day, PACE’s observers remained at the polling station after voting was completed to monitor the closing and counting process, and until all the counted ballots had been transported to the township tabulation centers. Overall, the closing and counting process was transparent and accountable. The process was open to the public, the counting process was conducted in a consistent manner, and the procedures necessary for the closing were taken according to UEC guidelines.

The level of transparency and accountability of the tabulation process could be measured by assessing whether the results from the polling station are tabulated correctly, the process is open to the public and all the regulations are followed. After polling stations closed on April 1, PACE deployed observers to all 22 townships election sub-commissions to monitor the tabulation of results.

Generally, PACE found that the process was open to the public and all observers were allowed to monitor the whole process with no restrictions. All necessary procedures were followed at nearly all tabulation centers. At all counting centers, any advance votes were received by 4 pm on election-day. Additional findings include:

  • PACE observers were allowed to monitor the tabulation process in all townships without restrictions.
  • There were party and candidate agents present at all sub-commissions. NLD agents were present at 21 of the 22 tabulation centers, USDP at 20, other Burman parties at 10, ethnic parties at 8, and agents for independent candidates at 5.
  • The sub-commissions did not accept any advance votes after the 4pm deadline.
  • PACE observers reported being able to see the marks on the advance voting ballots in 21 of the 22 tabulation centers.
  • All sensitive materials, such as ballots and results forms, were stored securely in 20 of the 22 tabulation centers.
  • PACE observers did not witness any instances of interference, intimidation or harassment in the tabulation process.
  • Party agents did not submit any complaints at 20 of the 22 tabulation centers. USDP submitted complaints at 2 centers, and NLD in 1.
  • The tabulation process was completed on April 1 in 13 of the 22 townships. The rest of the centers completed the process on April 2.