Survey findings

The data underlying this analysis of Burmese migrant worker remittances are drawn from a survey of about 1000 such workers undertaken by the authors in 2002–03. The survey was conducted in 12 provinces of Thailand, selected according to their relative importance as areas of settlement and employment of Burmese.[15] By far the majority of the workers interviewed had fled the violent repression in Burma that followed the uprisings of 1988, with most arriving during the latter half of the 1990s. The survey revealed that prominent in the list of factors driving the flight of the Burmese, in addition to fleeing the country’s political repression and civil wars, was Burma’s grinding poverty and lack of economic opportunities. Not unexpectedly, given their proximity to Thailand and the long experience of conflict in these areas, the majority of the workers surveyed (and Burmese in Thailand generally) were from rural areas of Burma’s Karen, Mon and Shan States, as well from Tenasserim Division. The median income for Burmese migrant workers employed in their first job in Thailand was about 2500 baht (Bt) a month. Only 40 per cent of the workers surveyed were in possession of a formal work permit conferring legal employment status. Workers holding work permits were over-represented in jobs earning more than Bt5000 a month, while the converse was true for those earning below Bt1000. The workers in the survey were employed across a number of industries, including fishing, construction, retail trade, hotels and restaurants, household service, food processing, agriculture, forestry and quarrying. Incomes earned varied widely across these industries, the highest paid being those engaged in quarrying and the lowest in agriculture and forestry. A majority (62 per cent) of those surveyed were men.

Results: amounts sent

As can be seen from Table 5.1, the median amount sent home by the 524 survey recipients who declared making remittance payments was Bt15 000 (about $US575). This is an annual aggregate figure, nominated by the respondents as the total they sent home for the previous 12 months. The maximum amount sent by any single worker was Bt3 million, while the lowest non-zero annual remittance total was Bt3000.

Table 5.1 Summary statistics for remittances (n = 524/1023)

Summary measure

Value (baht)

Sum

10 034 083

Mean

19 149

Median

15 000

Mode

10 000

Aggregate ratio to income

0.38

The aggregate ratio of annual remittance payments to annual disposable income of 38 per cent is quite high relative to experiences elsewhere. Moreover, the ratio is almost certainly understated since the survey includes bands of income rather than discrete amounts and the calculation here includes income at the top of each nominated band. Estimates for the United States suggest that migrant workers send home between 20 and 40 per cent of their aggregate earnings, but the concentration in most studies is between 20 and 25 per cent (Orozco 2006; Amuedo-Dorantes et al. 2005).

Should the survey results be anywhere near representative of the million or so Burmese in Thailand, the aggregate annual flows of remittances to Burma in 2002/03 from this source would have been in the order of $US300 million.[16] Such flows are nearly five times Burma’s ‘official’ remittance payments for the year, more than twice the amount of FDI received and they would represent about 5 per cent of Burma’s GDP.[17] Of course, these numbers are very rough, but they give some idea of the magnitude of the likely remittance payments to Burma and the potential they could yield for the country’s economy more broadly.

Not unexpectedly, individual remittances from migrant workers tend to rise with income (de la Garza and Lowell 2004). As can be seen in Figure 5.1, the survey data are consistent with this general rule. What appear to be strong falls at various income categories in the graph is largely a statistical anomaly—due to greater sampling variability caused by small numbers of respondents at these incomes.

Figure 5.1 Remittances by income group
Figure 5.1 Remittances by income group

Figure 5.2 Remittances by duration of residence in Thailand

Figure 5.2 Remittances by duration of residence in Thailand

A phenomenon identified in other places and contexts is that individual remittances decline with time, usually as a consequence of migrants putting down roots in their host country and losing touch with their former home. As Figure 5.2 indicates, such a decline in remittances is evident from our data too. The authors are, however, reluctant to draw the conventional inference since few Burmese establish de jure permanent settlement in Thailand (if not de facto residence for considerable periods). It is possible, for instance, that in Burma’s circumstances the depicted decline in remittances for individuals across time is less a function of them establishing themselves in Thailand than the fact that family members might themselves have subsequently left Burma. In this case, there could be no-one left to send remittances to.

Uses for remittances

What are Burmese migrant-worker remittances used for? Migrant-worker remittances everywhere are made and subsequently spent according to a hierarchy of needs. According to this survey, the remittances Burmese workers send from Thailand are used overwhelmingly to assist their families in basic survival. Some 96 per cent of respondents nominated this as their first order of priority. Indeed, many nominated family survival as their only motivation, with some taking the opportunity to annotate on our survey documents the living conditions faced by their families back in Burma.

As can be seen from Figure 5.3, other purposes (nominated on the survey itself according to priorities common in other contexts and circumstances) were negligible motivating factors. Such motivations, however minor, included (in order of importance assigned by survey respondents): to purchase or develop farm land; to establish a business; to meet education expenses; to repay debt; to hire workers in Burma; and to purchase consumer goods.

Figure 5.3 Remittances by intended recipient use
Figure 5.3 Remittances by intended recipient use

Figure 5.4 Remittance share by transfer method

Figure 5.4 Remittance share by transfer method

Methods and instruments

The most noteworthy fact revealed by the survey of remittance methods and instruments is the overwhelming dominance of informal funds transfers. As demonstrated in Figure 5.4, 94 per cent of respondents in the survey reported that they made their remittance payments via informal funds transfers. Within informal funds transfers, hundi was the dominant device (twice as large as any other), but significant proportions were simply carried into Burma by hand by either close family members or carriers and merchants. The nearly one-quarter share taken up by carriers and merchants could reflect a number of factors: most obviously, the porous nature of the Thai–Burma border, but also a possible remittance role played by brokers who bring many Burmese into Thailand in the first place. The small share (6 per cent) of migrant workers who carry their payments back into Burma themselves likely reflects the fact that Burma’s political situation dictates that few migrants are in a position to return home.

The dominance of hundi is not unexpected since, as alluded to above, it is a device particularly well suited to the situation faced by Burmese migrant workers. Using banks is almost impossible for most Burmese in Thailand because of the level of documentation required (as noted earlier, to use banks in Thailand, foreigners must establish a ‘non-resident’ baht bank account, which in turn requires a passport, a long-term visa or work permit, as well as letters of recommendation). Meanwhile, Burma’s political situation means that other formal transfer mechanisms (such as money-transfer firms) are not functional in the trade. Accordingly, and unless funds are physically carried across the border, hundi is the device preferred by most. As noted earlier, hundi relies on trust. Significantly then, the survey revealed no cases of fraud and numerous anecdotes of hundi dealer selection being based on positive word of mouth, reinforced by subsequent repeated dealings.

Other reasons for the dominance of hundi almost certainly include historical familiarity and cost. Hundi systems have a long history in Burma and in the colonial era they were championed by the Chettiar moneylenders, who, for nearly a century, were the country’s principal financiers. The Chettiars used hundi systems not only to transfer funds (domestically as well as internationally), but to provide a (critically important) credit instrument to Burmese cultivators. In this context, the Chettiars provided credit simply by making payments in advance of the hundi sender’s remittance—turning a hundi into something akin to the ‘bill of exchange’ long familiar in the West. Just like bills of exchange elsewhere, Western banks operating in Burma even ‘discounted’ hundis (that is, buying them at less than face value and thus providing the seller with finance at a discount ‘rate’ equivalent to an interest charge), thus connecting colonial Burma’s informal financial system with that of its formal and international equivalent—and providing Burma with the finance that went towards the transformation of the Irrawaddy (Ayeyarwady) Delta into the largest rice-growing region in the world. Eventually, in its perennial quest to create a viable tax system in Burma, the colonial government even supplied a specially authorised hundi ‘chapter’, upon which was payable a stamp duty. The Chettiars were effectively expelled from Burma post independence, but the hundi lived on as a principal device for trade throughout the country and beyond it. With the Chettiars gone, hundi in Burma ceased to be identified with any particular ethnic group. This legacy today separates the experience of hundi in Burma from those of many other parts of South-East Asia, where the trade is greatly dominated by ethnic Chinese and their businesses.[18]

While precision in this context is not remotely possible, the authors estimate that about 40 per cent of all of Burma’s ‘legal’ border trade is conducted via hundi, while more than 90 per cent of strictly illegal business (smuggling notably) is undertaken through hundi channels. Following the emergence of private banks in Burma in the early 1990s, hundi entered a period of decline as a remittance and credit device in the country. It has re-emerged strongly since the collapse of the most important of these banks in the financial crisis of 2002–03.

The survey respondents reported that the average cost of sending a hundi amounted to about 5 to 10 per cent of the principal sent (levied mostly via a discount/premium on the baht/kyat exchange rate). As noted earlier in this chapter, formal money-transfer firms typically charge between 10 and 15 per cent.

The small (6 per cent) share of remittances sent into Burma via formal banks is confined almost exclusively to migrant workers sending funds to Putao in Kachin State. Putao is a hill town that can be reached only by air. This, coupled with the fact that it is also home to four battalions of the Tatmadaw (army), conceivably makes the physical carrying of funds from Thailand, and that sent via hundi, highly vulnerable to ‘official’ expropriation.

Other regional variations of remittance method are also apparent in our survey data. The physical carrying of funds, for instance, is the overwhelming practice of those sending funds to Shan State. In contrast, almost 100 per cent of workers remitting funds to Tenasserim (Tanyintharyi) and Mon State use hundi.

Hundi is the dominant remittance device according to the number of Burmese migrant workers using it, but it is also the favoured instrument according to value. In this context, hundi’s lead over other mechanisms is marginal, and to some extent the most remarkable fact observable in Figure 5.5 is the high degree of equality of the various instruments by value (the numerals written in each of the columns below represent the raw numbers of individuals nominating each of the remittance devices).

Figure 5.5 Remittance magnitudes by method
Figure 5.5 Remittance magnitudes by method

Gender differences

This survey yields some interesting differences in remittance patterns according to gender, of which the essential finding is that female Burmese migrant workers send home a higher proportion of their income (40 per cent of the maximum) than males (36 per cent). Male workers earn more, but their median annual remittance of Bt15 000 is identical to that of female workers. This finding is similar to that of other studies exploring gender differences in remittance patterns in other country pairs (see, for instance, Moreno 2005). The differences are more pronounced among lower-income workers (of which women constitute a larger proportion than the population generally) than among those earning higher incomes. Curiously, the survey also reveals differences according to gender in terms of remittance methods and instruments. Women are more likely (18 per cent) to use a bank to send their money home (though the absolute numbers here remain small) and are more likely to send money by hand through another person (25 per cent), but are less likely to use hundi (13 per cent).

Table 5.2 Remittances and gender

Summary measure

Female

Male

Sum

Bt3 869 100

Bt6 164 983

Mean

Bt18 966

Bt19 488

Median

Bt15 000

Bt15 000

Mode

Bt10 000

Bt10 000

Coefficient of variation

73.9%

112.1%

Aggregate ratio to income (2003 maximum)

0.40

0.36




[15] For a comprehensive outline of the survey itself, its methodology and the provinces selected, see Vicary (2004) and Bradford and Vicary (2005).

[16] We have assumed an identical ratio of remittance senders to total migrant population identified by our survey. Of course, the figure of ‘one million’ Burmese in Thailand is not much more than an informed ‘guestimate’—albeit, we expect, a conservative one.

[17] Investment and GDP data are based on that in IMF (2007).

[18] See, for instance, Bolt (2000) for one of but many studies investigating this phenomenon.