Determinants of migration streams in Myanmar Nyi Nyi and Phillip - - PDF document

Determinants of migration streams in Myanmar

Nyi Nyi and Phillip Guest September, 2017

A scientific paper presented at the 2017 International Population Conference of the International Union for the Scientific Study of Population (IUSSP) in Cape Town, South Africa at the Cape Town International Conference Centre (CTICC) from 29 October to 4 November 2017.

2

Determinants of migration streams in Myanmar

Introduction The 2014 Myanmar Population and Housing Census is the first census to be undertaken in the country since 1983. Findings from this census suggest that the level of internal migration in Myanmar is similar to that of neighbouring countries (Bell and Charles-Edwards, 2013). Many of the differences between countries can be explained by the level of aggregation used in defining

• migration. For Myanmar, migration is defined as moving between townships (a relatively small

administrative unit) which leads to a relatively high level of movement. Over the lifetime of individuals, 19.3 per cent reported moving at least once. For internal migration within the five-year period before the Census, 7 per cent reported moving (Department of Population, 2016). Before the 2014 Census there had been limited research on internal migration or urbanization within Myanmar (Department of Population, 2013) although the demographic structure of the country had been studied. The country is predominately rural, with only 29.6 percent of the population residing in urban areas. This is the second lowest level of urbanization in Southeast Asia (UNESCAP, 2014). Yangon, with a population of 5.2 million, is the most urbanized region of the country, accounting for about 35 per cent of the total urban population, although it contains approximately 40 percent of migrants measured in the 2014 Census. It is projected that by 2040, the city will become a megacity with a population of 10 million, catching up with the trend of the rest of the Southeast Asian nations (JICA, 2013). With the recent opening up of the country, it is also expected that urbanization will take place at an increasingly rapid rate across many cities and towns of Myanmar. This increasing tempo of urbanization was expected to be fed by increasing rural to urban migration. The definition of migration used in the Census is designed to capture permanent or semi- permanent changes of residence. The criterion of six months used to establish the time spent in their usual residence results in those who move on a temporary basis of less than six months not being included in the definition of migration. Temporary forms of migration are typically the predominant form of movement in Southeast Asian countries (Hugo, 2012) and the exclusion of this type of mobility will result in estimates of the level of migration recorded from the Census as being too low, particularly the movement for agricultural workers which are generally seasonal in nature (Mahajarn and Myint, 2015). Differentials in the characteristics of migrants, who move for a short period of time, often circulating between areas of origin and destination, are also likely to differ. For example, Guest (1999) has shown for Viet Nam, temporary migrants are likely to be older, married and have less education than migrants who move for longer periods. However, the inclusion of questions in a census questionnaire that would allow temporary migrants to be identified is not feasible given the dual needs to include questions that measure a range of characteristics and events while also restricting the length of the questionnaire. Hence, censuses worldwide focus on movement of a longer duration (Deshingkar and Grimm, 2005).

3

Migration is measured in the 2014 Census as movement between townships. Myanmar is comprised of 414 townships (including 84 sub-townships). Movement within a township was not

• asked. Much of this movement likely involves changes of place of residence rather than a change
• f social networks. These may occur over a very short distance and without disrupting social

patterns, although some moves may take place over a longer distance as townships range in size from 1.6 square kilometer for Latha Township in Yangon West district to 11,452 square kilometers for Tanintharyi Township in Myeik District. Internationally there is considerable variation in the number of administrative districts used to define migration: in Southeast Asia in the 2010 round of censuses Indonesia used 33 provinces to define migration while Vietnam used 9,111 administrative units to define migration (Bell and Charles-Edwards, 2013). One of the most striking findings from the Census is the direction of the flow of migration. Almost half of recent migration (that which occurred within the five years prior to the Census)

• ccurred between urban areas and less than 10 per cent of movement was from rural to urban
• areas. Meanwhile, migration from rural areas was directed primarily towards other rural areas.

These patterns are unusual when viewed from a regional or international perspective. A population that is predominately rural would be expected to have levels of rural to urban flows that were approaching the flows of rural to rural movement. In addition, it was found that there were more inter-state/region movements than intra-state/region movements. Another interesting finding was that a large proportion of movement within Myanmar revolved around Yangon, either as movement into Yangon or movement among districts Yangon. Thus, the purpose of this study if to examine the determinants of migration streams in Myanmar. Data and research methods The data used in this paper comes from recent surveys conducted by the Department of Population and the 2014 Census. The methodology used in the Census has been described elsewhere (Department of Population 2015). The total number of persons enumerated in the Census was 50,279,900. It is estimated that 1,206,400 were not enumerated. The analysis focuses

• n individuals who are living in what is called in the 2014 Census, ‘conventional’ households. This

analysis does not include the 2,349,901 persons who were enumerated as living in institutions. These persons were not asked the questions that are used to determine the migration status of an

• individual. It can be assumed, however, that the levels of mobility are likely to have been high

among some segments of the institutional population. The analysis concentrates on the following four major streams of migration: rural-rural, rural-urban, urban-rural and urban-urban. International comparisons of levels of urbanization are hampered by variations in the definitions of urban and rural areas. There is no consensus as to how an administrative area is classified as urban or rural. The 2014 Myanmar Population and Housing Census adopted the designations employed by the General Administration Department (GAD) of the Ministry of Home Affairs in designating lower level administrative areas as urban (wards) or rural areas (villages).

4

Drawing upon interviews with officials from the GAD, the UN Habitat states that there were 288 urban centers under the Towns Act and the Municipal Act (UN-Habitat, 1991). The Towns Act may apply to centers of population less than 10,000, “with densities and functions of a sufficiently urban nature for urban wards (as opposed to rural village tracts)” (UN-Habitat, 1991:53), whereas municipalities are (were) defined as centers with over 10,000 population. In addition to property- based taxation, other urban functions—from planning to implementation and management of physical and social services—were also described as criteria for being “urban”. Consequently, other than the population within some well-defined administrative boundaries such as Yangon City, it is more challenging to identify urban centers or towns from the 2014 Census data. Results Estimates of the share of the four major migration streams are shown in the figure below (Figure 1). These are defined as the four flows that occur between rural and urban areas (that is, urban-urban, urban-rural, rural-urban and rural-rural). The figure displays lifetime migration rather than five-year migration rates because the earliest points in the data series are only available for lifetime migration. There was a large increase in the proportion of lifetime migrants moving between urban areas over the 23-year time period being considered. At the time of the 2014 Census, almost 47 per cent of migrants were classified as having engaged in this type of movement. Rural-to-rural migration comprised the second largest stream accounting for almost 30 per cent of

• migrants. The only stream that had declined over the whole period was the rural to urban stream,

which accounted for less than 10 per cent of migrants at the time of the 2014 Census. Figure 1. Estimates of percentage of lifetime migration streams from surveys and the 2014 Census Source: Data for 1991 PCFS, 2001 FHRS and 2007 FHRS are from DoP (2013).

38.6 33.5 40.5 46.9 12.9 9.1 9.2 14.16 30.4 25.4 24.7 9.53 18.1 32 25.6 29.04

5 10 15 20 25 30 35 40 45 50 1991 PCFS 2001 FHRS 2007 FHRS 2014 Census

% Lifetime Migration

Urban - Urban Urban - Rural Rural - Urban Rural - Rural

5

In Table 1 the 2014 Census data is used to display data for five-year migration streams by

• sex. The table shows that 3.3 per cent of all persons in conventional households had migrated from

an urban to an urban area in the five years prior to the Census (this equates to almost half (47.3 per cent) of all migration streams), while 1.1 per cent, 0.7 per cent and 1.8 per cent migrated from urban to rural, rural to urban and rural to rural areas respectively (equating, to 16.0 per cent, 10.4 per cent and 25.7 per cent respectively of all moves). There was very little difference between the migration patterns of males and females. Table 1. Migration streams for recent migrants in conventional households by sex, 2014 Census

Sex Migration stream Non- migrant Total Urban- Urban Urban- Rural Rural- Urban Rural- Rural Migrant from urban areas* Migrant from rural areas* Total 1,587,121 537,762 348,691 863,419 13,971 8,378 44,559,183 47,918,525 3.3% 1.1% 0.7% 1.8% 0.0% 0.0% 93.0% 100.0% Males 742,389 254,116 173,161 429,070 6,437 3,973 20,939,382 22,548,528 3.3% 1.1% 0.8% 1.9% 0.0% 0.0% 92.9% 100.0% Female 844,732 283,646 175,530 434,349 7,534 4,405 23,619,801 25,369,997 3.3% 1.1% 0.7% 1.7% 0.0% 0.0% 93.1% 100.0%

* Migrants whose current place of usual residence (which may have been different from where they were enumerated) was not recorded. A proxy measure of distance is the type of migration stream recorded (between townships within districts, between districts within states/regions and between states/regions) . It is assumed that the level of disrupture to social ties increases from migration that occurs between townships within districts to migration that occurs between states/regions. Migration streams between Townships, Districts and State/Region are shown for recent migration in Table 2. Of the total population in conventional households, 93 per cent had not migrated in the five years prior to the Census. Some 2.7 per cent had migrated between Townships within Districts, 1.1 per cent had moved between Districts but within States/ Regions and the remaining 3.2 per cent had migrated between States/Regions. The amount of movement that occurs within districts is only marginally below the level that

• ccurs between states/regions. Therefore, the highest level of migration flows is between

states/regions it is followed closely by relatively short-distance movement within districts.

6

Table 2. Migration patterns for recent migrants in conventional households by sex, 2014 Census

Sex Recent migration Total Between Townships within Districts Between Districts within States/Regions Between States/Regions Non- migrant Total 1,291,185 524,736 1,543,421 44,559,183 47,918,525 2.7% 1.1% 3.2% 93.0% 100.0% Male 612,803 251,660 744,683 20,939,382 22,548,528 2.7% 1.1% 3.3% 92.9% 100.0% Female 678,382 273,076 798,738 23,619,801 25,369,997 2.7% 1.1% 3.1% 93.1% 100.0%

Table 3 displays the 20 most numerous flows of inter-district population movement. Of the top 20 flows, all, except one, has one of the four districts that comprise Yangon as the destination. North Yangon is the main destination of the flows, appearing 12 times in the top 20. North Yangon has expanded its industrial base over the last decade and many persons appear to move to the District to work in the industrial sector. In 2011, many of the 23 Industrial Zones in Yangon were in North Yangon. Industrial employment almost tripled in the two decades from 1988 with most of this growth occurring in Yangon (Zaw and Kudo, 2011). Also, the central business district (CBD) is located in West Yangon but rents have risen rapidly in recent years forcing many people to leave for the outer areas of the city. East Yangon is less crowded and has lower rents and this district appears to be a beneficiary of this movement. Analysis of the industry data from the Census indicate that the labor force in North Yangon contains the highest number of manufacturing sector

• jobs. This is especially so for female migrants where almost 50 percent of the labor force is

employed in manufacturing. This is followed by East Yangon district where almost one quarter of females is employed in manufacturing. There are also industrial zones that have been established outside of Yangon. Mandalay has four zones, Ayeyawady has three, Bago and Magway have two each, and Mon, Sagaing, Shan and Tanintharyi have one each. Each of these zones are designed specifically to attract local, and in some instances foreign investment. All are attracting workers and are contributing to increased migration.

7

Table 3. Top 20 District-to-District flows for recent migrants, 2014 Census Rank District to District flow Number of migrants 1 West Yangon to East Yangon 56,601 2 Phayapon to North Yangon 51,086 3 West Yangon to North Yangon 43,044 4 Hinthada to North Yangon 35,311 5 Maubin to North Yangon 33,369 6 Labutta to North Yangon 29,796 7 Thayawady to North Yangon 29,362 8 North Yangon to East Yangon 28,368 9 Pathein to North Yangon 27,852 10 East Yangon to West Yangon 26,257 11 East Yangon to North Yangon 25,652 12 Phayapon to East Yangon 24,891 13 South Yangon to North Yangon 24,660 14 Myingyan to Mandalay 24,500 15 Myaungnya to North Yangon 21,694 16 Bago to East Yangon 20,807 17 Bago to North Yangon 18,663 18 North Yangon to West Yangon 17,519 19 Magway to North Yangon 16,232 20 Pathein to East Yangon 14,835 Table 4 shows that migrants who moved from an urban place to another urban place had a much higher level of completed education than any of the other migration streams. The differences

8

are large. Compared with the 44.7 per cent of urban to urban migrants who completed high school

• r above (including vocational training), the corresponding proportions for the three other

migration streams were 23.3 per cent for urban-to-rural, 31.4 per cent for rural-to-urban, and 14.3 per cent for rural-to-rural migrants. Therefore, it can be concluded that the dominant position of urban to urban migration, compared to other migration streams, is partly a function of relatively high levels of education of the participants of this stream. Table 4. Highest level of education completed of recent migrants aged five and over by migration by migration stream, 2014 Census Highest level of educational attainment Migration streams Urban-Urban Urban-Rural Rural-Urban Rural-Rural None 72,353 38,221 24,932 112,116 4.8% 7.4% 7.9% 14.2% Primary 382,125 210,681 106,287 380,472 25.5% 41.0% 33.5% 48.1% Middle school 371,151 142,084 84,922 177,954 24.8% 27.7% 26.7% 22.5% High school 301,862 67,665 54,879 69,759 20.2% 13.2% 17.3% 8.8% Diploma 7,858 1,261 1,685 1,696 0.5% 0.2% 0.5% 0.2% College or University 334,819 48,634 40,028 39,884 22.4% 9.5% 12.6% 5.0% Post-graduate 18,954 1,453 2,500 1,518 1.3% 0.3% 0.8% 0.2% Vocational training 4,420 723 777 837 0.3% 0.1% 0.2% 0.1% Other 3,920 2,531 1,464 7,230 0.3% 0.5% 0.5% 0.9% TOTAL 1,497,462 513,253 317,474 791,466 100.0% 100.0% 100.0% 100.0% Table 5 provides the odd ratios from a multinomial logistic regression of the probability of participating in a specified migration stream. The omitted category for the dependent variable is

9

urban to urban migration. As expected, an increase of one year in age is associated with a reduction in the odds of being categorized in any of the three migration streams compared to being an urban to urban migrant. The non-linear relationship between age and migration probabilities is modeled by including a term for age squared into the equation. Of the four variables measured at the township level (rather than at the individual level), increases in the percent urban and the percent employed in agriculture at the township level reduce the odds of being an urban-rural, rural-urban or rural-rural migrant compared to being an urban-urban migrant. Therefore, higher percentages of urban population and higher percentages of the population employed in agriculture in the destination township are both associated with the

• dds of higher urban to urban migration. The percent of the population with a middle school or

higher education and the net migration rate of the township have only a limited impact on the odds

• f being an urban to urban migrant compared to the other three migration streams.

After controlling for other variables, the odds of a male being an urban-urban migrant, compared to being in one of the other three migration streams is higher than that of a female. Males are less likely than females to move between rural townships, from rural to urban townships and from urban to rural townships. The odds are particularly lower for males, compared to females, for movement from rural to rural townships compared to urban to urban townships. The relationship of a person in the household only appears to be associated with the probability of being an urban-urban migrant when we compare it with being a rural-urban migrant. Those persons in the household who are not household heads have odds of being a rural-urban migrant compared to an urban-urban migrant that are 16 percent lower than the odds of the household head. Being never-married, however, is associated with much lower odds than other marital statuses of being a rural-urban migrant or a rural-rural migrant compared to being an urban-urban migrant. Persons with a high school or higher level of education, compared to less than a high school level of education, are much more likely to be urban-urban migrants compared to being in any of the other three migration streams. The comparison is most dramatic for rural-rural migrants, where a person with a lower level of education (below high school) has odds of being a rural to rural migrant, compared to an urban to urban migrant, which are 3.57 times higher than that of a person with a high school or higher level of education. Occupation is measured at the time of the Census and thus is not ideal for analyzing the determinants of migration streams since occupation is measured after migration. Being currently in an agricultural occupation, compared to not being in the labour force is associated with much high

• dds of being in any the three migration streams, particularly a rural to rural migrant, compared to

being an urban to urban migrant. The other effects are more limited, with being in a modern sector

• ccupation, compared to not being in the labour force, being associated with lower odds of being a

rural to rural migrant and a rural to urban migrant compared to being an urban to urban migrant.

10

Table 5. Multinomial logistic regression estimates of the odds of undertaking specified stream of internal migration Variable Migration stream Urban-Rural Rural-Urban Rural-Rural Age 0.960 0.991 0.953 Age squared 1.000 1.000 1.000 Net Migration 1.000 1.000 1.000 Percentage middle school and above 1.000 1.000 1.000 Percentage urban 1.000 0.944 0.946 Percentage in agriculture 0.990 0.948 0.970 Sex Male 0.894 0.954 0.877 Female 1.000 1.000 1.000 Head of Household No 0.987 0.835 0.949 Yes 1.000 1.000 1.000 Education Below high school 2.475 1.478 3.570 High school and above 1.000 1.000 1.000 Never Married Other 0.935 1.330 1.593 Never married 1.000 1.000 1.000 Occupation Modern sector occupation 1.132 0.883 0.701 Agricultural occupation 2.075 3.746 8.880 Other occupation 1.569 1.194 1.724 Not in labour force 1.000 1.000 1.000 Residence before migration Outside of Yangon 5.045 0.904 6.544 Resident in Yangon 1.000 1.000 1.000 N=2612519 Note: All variables are significant at the 0.000 level. The analysis for this table is restricted to persons aged 15 and above.

11

The variable measuring the place of residence before migration (Yangon or outside of Yangon) is strongly associated with urban to rural migration and rural to rural migration for those who lived outside of Yangon. For those persons living outside of Yangon, compared to those who were residents of Yangon, there was a 10 percent reduction in the odds of rural to urban migration compared to urban-urban migration. In other words, residents of Yangon were more likely to engage in urban to urban migration compared to residents from outside of Yangon. In Table 6 we show predicted probabilities of being in each of the four migration streams. These are shown for the variables sex, education and place of previous residence. The predicted probabilities are based on the full model outlined in Table 5. Table 6. Predicted probabilities of migration in each of the migration streams for specified variables Variables Migration stream Urban-Urban Urban-Rural Rural-Urban Rural-Rural Sex Male 0.4762 0.1645 0.1028 0.2565 Female 0.5019 0.1690 0.0949 0.2342 Education Below High school 0.3820 0.1956 0.0980 0.3244 High school and above 0.6675 0.1195 0.0966 0.1134 Residence before migration Outside of Yangon 0.3879 0.1976 0.1025 0.3120 Yangon 0.8159 0.0685 0.0861 0.0295 The predicted probabilities illustrate the dominance of urban to urban migration for those with a higher level of education. While, almost 67 percent of those with a high school or above level of education are in the urban-urban category, with only 12 percent urban-rural migrants, 10 percent rural-urban migrants and 11 percent rural-rural migrants. For those with a below high school level of education, only 38 percent were urban-urban migrants. Similarly, if a migrant was resident in Yangon before migration, almost 82 percent were predicted to be urban-urban migrants compared to the 39 percent of migrants who lived outside of Yangon. The differences by sex are much less, with 48 percent of males and 50 percent of females being classified as urban-urban migrants. Discussion and conclusion The analysis of internal migration patterns within Myanmar clearly shows many of the expected patterns. Migration flows are directed primarily to places where economic opportunities are the greatest. This includes Yangon, which contains slightly over 40 percent of all migrants, but also involves other urban places. Typical migrants are young and relatively well educated. Females

12

are more likely to migrate than males. And employment in the industrial sector of the economy is much more common for migrants than for non-migrants. There were also some findings that were not expected. Chief among these were the direction of the flows of migration. Almost half of recent migration occurred between urban areas, and less than 10 per cent of movement was from rural to urban areas. While the definition of ‘migration’ employed in the Census undoubtedly resulted in many moves from rural areas not being included as migrations, the results do suggest that for more permanent migration the flows are predominately urban to urban. More permanent migration from rural areas was directed towards

• ther rural areas. The continuing importance of rural to rural migration as a major stream of

migration both reflects the large rural population in Myanmar and the availability of work in mining, much of it involving low-skilled occupations, and in certain types of agriculture. Migrants tend to be related to the head of the households into which they move. This points to the importance of social networks in facilitating movement and their role in assisting in the settlement of migrants, and might help explain the lack of rural to urban flows of migrants. The numerical dominance of urban to urban migration streams is facilitated by the dominance of flows

• f migration between townships, and between districts, within Yangon.

A large portion of this movement seems to be a result of the emphasis that Yangon has received as the hub of development in Myanmar. This development has led to the expansion of the city with many of the occupational opportunities being located away from the heart of the city. Some of this development has resulted from explicit policy decisions of the government. For example, the opening up of many industrial zones in outlying areas has both created much sought- after employment in the enterprises that have located in these zones, but have also created

• pportunities for those who provide services in these new areas. This has resulted in a shift of

population from inner city areas to areas located on the outskirts. Some of this shift has occurred for those persons who are more highly educated and skilled and hence were more likely to be living in Yangon than in other states or regions. An analysis of the characteristics of both individuals and townships associated with different migration streams provides indirect evidence for the processes that have been suggested to be

• perating. Urban to unban migration is more likely to be undertaken by older migrants than the
• ther three streams of migration. This probably reflects the higher skill requirements of

employment within the urban to urban migration stream. After controlling for other variables in the model, females, compared to males, are more likely to be involved in rural to rural migration than urban to urban migration. Some of the movement to rural areas for females likely involves permanent relocation due to marriage. This is reflected in the finding that rural to rural migrants are much less likely to be never married than are migrants involved in urban to urban migration. Township characteristics, particularly the percent of the township population that was urban and the percent of township population that were employed in agriculture were related to higher probabilities of participation in urban to urban migration. While this is understandable in terms of the percent of the population in urban areas in a township the relationship between

13

agriculture and participation in urban to urban migration requires greater study. It is possible that this may be a result of industrial zones being located in townships that still have a strong agricultural sector. One finding that stand out in the analysis is the importance of education in defining the migration stream of individuals. Migrants with a lower level of education (below high school) were more likely to be involved in rural to urban, urban to rural and rural to rural migration streams compared to those persons who moved between urban areas. In particular, lower education was associated with urban to rural migration (probably involving return migration) and rural to rural

• migration. Higher education is probably required to access many new opportunities in urban areas.

Although the current occupation was used in the analysis as the occupation before migration was not available, the results indicate that those who are working in agricultural

• ccupations are much less likely to engage in urban to urban migration and much more likely to be

involved in rural to rural migration. The results show that while the migration patterns of Myanmar are different from those of many other countries they can be understood within the context of the development patterns of the country. High levels of investment in one city – Yangon – have created situations where new employment opportunities are being concentrated primarily in Yangon, and many of these

• pportunities are located in districts away from the center of the city causing relatively high levels
• f movement from inner city areas to townships on the periphery of the city. On a much smaller

scale, it appears that in Mandalay and other large cities a similar process is operating. Those who are more highly educated, and presumably more highly skilled, are taking part in this migration. Meanwhile, the large rural population provides individuals with a low level of education the

• pportunity to move within rural areas to access mining and agricultural jobs.

References Bell, Martin and Elin Charles-Edwards (2013). Cross-National Comparisons of Internal Migration: An Update on Global Patterns and Trends, Technical Paper No. 2013/1, Population Division, United Nations. Department of Population and UNFPA (2013). Levels, Trends and Patterns of Internal Migration in Myanmar, Ministry of Immigration and Population/UNFPA, September 2013. Department of Population (2015). The 2014 Myanmar Population and Housing Census, Union Report, Ministry of Immigration and Population, May 2015. Department of Population (2016). The 2014 Myanmar Population and Housing Census, Thematic Report on Migration and Urbanization, Ministry of Labour, Immigration and Population, Nay Pyi Taw, Myanmar.

14

Deshingkar, Priya and Sven Grimm (2005). Internal Migration and Development: A Global Perspective, IOM Migration Research Series No.19, International Organization for Migration, Geneva. Guest, Philip (1999). The Dynamics of Internal Migration in Vietnam. UNDP Discussion Paper No. 1, UNDP, Hanoi, Vietnam. Hugo, Graeme (2012). “Changing Patterns of Population Mobility in Southeast Asia”, in Lindy Williams and Michael Guest (Eds.), Demographic Change in Southeast Asia. SEAP, Cornell University, Ithaca, New York, pp.121-163. JICA (2013). The Project for the Strategic Urban Development Plan of the Greater Yangon, Final Report 1, Part-I: The Current Conditions, Japan International Cooperation Agency, The Republic of Union of Myanmar. Maharjan, Amina and Theingi Myint (2015). Internal Migration Labour Migration Study in the Dry Zone, Shan State and the Southeast of Myanmar, HELVETAS Swiss Intercooperation Myanmar, February 2015. UNESCAP (2015). ESACP Statistical Database, ESCAP, Bangkok, Published

• nline

at hhtp/unescap.org/stat/data, accessed on 29 August, 2015. United Nations Habitat (2009). Better Information, Better Cities: Monitoring the Habitat Agenda and the Millennium Development Goals-Slums Target, United Nations Human Settlements Programme, Nairobi, Kenya. Zaw, Myinmo and Toshihiro Kudo (2011) “A study on economic corridors and industrial zones, ports and metropolitan and alternative roads in Myanmar”, in Intra- and Inter-City Connectivity in the Mekong Region, edited by Masami Ishida, BRC Research Report No. 6, Bangkok Research Center, IDE-JETRO, Bangkok, Thailand.