Chain Migration Through the Social Network: Experience of Labour Migrants in Kuwait

Published by Blackwell Publishers Ltd.,108 Cowley Road, Oxford OX4 1JF, UK, and350 Main Street, Malden, MA 02148, USA.

© 1999 IOMInternational Migration Vol. 37 (2) 1999

ISSN 0020-7985

Chain Migration Through the Social Network:Experience of Labour Migrants in Kuwait1

Nasra M. Shah* and Indu Menon*

ABSTRACT

Labour migration to the Gulf countries is predominantly contract based anda majority of workers fall below the salary ceiling necessary for sponsoringfamily members. Despite this, social networks have expanded in Kuwait,primarily in the form of sponsorship of additional labour migrants by thosealready in the country.

The objectives of the article are to describe how the process of arrangingsponsorship works, to delineate the predictors of moving through a friend orrelative, or arranging sponsorship for a subsequent labour migrant, and toassess the “multiplier” effect of the above process. The article is based on asurvey among 800 South Asian skilled and unskilled male migrants,200 each from Bangladesh, India, Pakistan and Sri Lanka.

About 34 per cent of all respondents moved through friends or relatives,50 per cent through agents and 16 per cent through direct hiring or thegovernment bureau. Logistic regression analysis indicated that the factorsmost likely to predict a move through the social network consisted of beinga Pakistani or Indian, being a Muslim, and possessing some skills.

One-quarter of all migrants had arranged the visa for another migrant sincethey came to Kuwait. Logistic regression analysis showed the respondent’sduration of stay in Kuwait to be an exceptionally important predictor forarranging sponsorship for an additional migrant.

Monthly income, being married and being a Muslim were also positivelyassociated with arranging sponsorship. In terms of the multiplier effect,about 0.78 visas had been arranged per migrant. Pakistanis, who had thelongest average duration of stay in Kuwait, arranged the maximum number(1.6) of visas on average.

* Department of Community Medicine, Kuwait University, Kuwait.

362 Shah and Menon

INTRODUCTION

Labour migration to the Gulf countries is usually described as a fairly wellorganized and well regulated process which occurs mainly through recruitmentagents or government agencies (Gunatilleke, 1991; Shah, 1994). The role ofinformal social channels consisting of relatives, friends or community membersfrom the home country is increasingly recognized as an important channel forencouraging additional migration and aiding in the adjustment and sustenanceof new migrants (Gunatilleke, 1998; Azam, 1998; Nair, 1998). Migrants to theGulf are employed on labour contracts and do not have the right to citizenshipor permanent residence. Only a small proportion of migrant workers are in jobswhich offer salaries that allow their families to accompany them. In order to beentitled to bring the family, the worker’s salary must be 450 Kuwait Dinar ifemployed in the public sector, and KD 650 if employed in the private sector(Kuwait Times, March 8, 1994). The majority of skilled and unskilled workersearn below these salary ceilings, and those who earn above the ceilings may beaccompanied only by wife and children under certain ages.

Restrictions on family migration have not, however, prevented the outwardexpansion of social networks through other forms of migration. Sponsorshipin this case has consisted mainly of other labour migrants who may beconnected to the sponsor through such ties as place of origin, family orfriendship. Sponsorship has a special meaning; arranging the migration of asubsequent worker, in some cases through buying a work visa. This type ofmigration has been referred to as “serial migration” by Banerjee (1983). In asurvey on which this article is based, it was found that visas had been arrangedby friends or relatives in about one-third of the respondents (Shah, 1998a).It was also found that those who moved through friends and relatives wererelatively more successful in terms of salary earned and being happierwith the job, compared with those who moved through recruitment agents(Shah, 1998b).

The objectives of the article are to describe how the process of sponsorshipworks, who has friends in the host country, who is most likely to move throughfriends, who is most likely to arrange migration for a friend or relative, andto assess the numerical implications of such sponsorship in terms of its“multiplier” effect.

RELATED PREVIOUS RESEARCH

Several studies on Asia and elsewhere have shown that personal networks offriends and relatives are crucial supporters of migration. A study of Bangladeshconcluded that about 63 per cent of respondents had received information about

363Chain migration through the social network

job prospects in the Middle East from friends and relatives, about 12 per centhad arranged the move through friends, and about 16 per cent had been helpedby friends in financing the cost of migration (Mahmood, 1991). In Kerala,India, Nair (1991, 1998) reported that a large majority of the returnees hadarranged Middle East jobs through friends and relatives, and about one-thirdhad received help from friends to finance the move. For Pakistan, Khan (1991)reported that relatives had played an important role in “pulling” otherrelatives: 42 per cent of the migrants had a relative already abroad when theymoved, and the move was arranged by friends or relatives abroad in the caseof about one-third of the migrants. Recent research from Pakistan (Azam,1998) supports these findings. Another study on Pakistan (Gilani et al., 1981),reported that about three-quarters of the migrants had obtained their MiddleEast jobs through friends and relatives. A study on Sri Lanka showed thatabout 88 per cent of male migrants arranged their employment through arecruiting agent, and only about 5 per cent had arranged it through a friend(Gunatilleke, 1991). Recent data from airport surveys suggest that the role offriends and relatives in arranging migration is much greater than formerlybelieved (Gunatilleke, 1994). The experience of Sri Lanka also suggests thatworkers moving through friends are in a more secure position than thosemoving through agents. The former also receive more reliable informationabout the job offer (Gunatilleke, 1998).

Research from countries in other parts of the world supports the salient roleplayed by personal networks in South Asia (Menon, 1988; Lindquist, 1993;Fuller et al., 1990; Singhanetra-Renard, 1992; Hugo, 1981, 1995; Banerjee, 1983;Boyd, 1989; Bocker, 1994). Family, friends and community members withmigration experience may provide initial contacts and information or actuallysponsor the move by a subsequent kin, friend or community member. Accordingto Hugo (1994), in an LDC “whenever a person migrates, every individual thatthey know acquires social capital in the form of a contact at the mover’sdestination”. Upon arrival in the host country, the migrant’s adjustment may begreatly facilitated by the presence of persons from the home country. In aFilipino fishing village, Lindquist (1993) found that networks were selectiveand opportunities were limited to those with the closest ties. The ones wholacked connection to the networks could still find overseas jobs, but they weremore likely to be faced with unreliable information, delays and unscrupulousrecruiting agents.

Moving through the informal network of friends and relatives results in severalfavourable consequences for the migrant. Massey and Gracia (1987) show thatin case of Mexican migrants to the US, the network lowers substantially thecosts and risk of international movement. Mexican migrants to the US withaccess to a personal US network had an average 4.4 point advantage in occupa-tional prestige scores over those without network access (Mullan, 1989).

364 Shah and Menon

Socio-occupational mobility of Mexican migrants was positively related withthe presence of social networks in the US and the utilisation of such networks.However, a down side of network support may be that new migrants becomerestricted to occupations similar to their friends/relatives with relativelylesser opportunity for occupational mobility (Banerjee, 1983; Caces, 1987;Pohjola, 1991).

In countries such as the US that allow the worker to bring in his family once hehas acquired permanent residence or citizenship, formal employment agreementsmay serve as a starting point that provide a bridgehead of migrants whorepresent one end of a migratory chain (Boyd, 1989). Massey (1986) arguesthat networks are key elements in understanding Mexican migration to the US.Expansion of networks leads to more migration, which leads to the expansionof the network. Changing economic circumstances and settling of the initialimmigrants can, however, result in a weakening of the migration chain, asshown in the case of Turkish migrants to the Netherlands (Bocker, 1994). Whenjobs became scarce, and the country’s immigration policy became morerestrictive, most settled Turkish migrants were less inclined to encouragekinsmen to move to the Netherlands. Settled Turkish migrants began to act asgatekeepers who prevented additional immigration instead of being bridgeheadsaiding additional immigration. An increase in the unemployment rate reducedthe inflow of labour migrants (Waldorf et al., 1990). At the same time, aid tonewcomers increased with the migrant’s stay but tended to decline as themigrant’s stay in the host country lengthened and ties with the home countryweakened.

The extent to which initial migrants create a multiplier effect has been a subjectof much debate in the US. Goering (1989) showed that the admission of onelegal immigrant could spawn the migration of 64 relatives. While there are manymethodological and technical difficulties in estimating the “true” multiplier,research indicates that early warnings about the explosive process of chainmigration were overstated. Jasso and Rosenweig (1986) calculated that everylegal immigrant admitted in a labour-certified category creates 0.6 to 0.7 newadult migrants and an additional 0.5 immigrant children within ten yearsof entry. Arnold et al. (1989) reported a projected multiplier of between 1.0 and1.8 relatives who might be expected to immigrate in future years from thePhilippines and between 0.5 and 0.7 from Korea. The projected multiplier wasfound to be much smaller than the theoretical multiplier because the latter takesinto account intentions to migrate, and adjusts for factors such as the potentialmigrant relative already being an immigrant or citizen. Information on themultiplier effect of South Asian migration is reported in only one studyavailable to us. Based on case studies of 18 Sri Lankan primary migrantsemployed for 1-8 years, Gunatilleke (1998) reported arrangements for theemployment of 154 others, resulting in a multiplier of 8.5.


Network formation is clearly a very important structural mechanism supportinginternational migration. Such networks play an especially significant role infacilitating illegal flows (Hugo, 1994), although the amount of informationavailable on the role of networks in the various stages of the move from SouthAsia to the Middle East is fairly limited, as is comparative information on theexperiences of migrants from various countries residing in the same hostcountry.

DATA

The data for this article were obtained during interviews in November-December, 1995 in Kuwait with 200 male migrant workers originating ineach of the four South Asian countries (Bangladesh, India, Pakistan, and SriLanka), resulting in a total sample of 800. The objective was to assess the roleof informal networks of friends and relatives in migration to Kuwait, and toanalyse the migration experience of workers moving through this channelcompared with those who had moved through recruitment agents, or had beenhired directly by an employer in the host country. The structured interviewswere supplemented by focus group discussions with migrants from eachcountry. Seven focus group sessions comprising 5-6 persons each wereconducted.

The study covered one district, Jaleeb Al-Shayoukh, which in mid-1995comprised about 6.6 per cent of Kuwait’s population. This district hada population of 123,874 persons, of whom 92 per cent were non-Kuwaitis.Asians comprised 67,867 (or 54.7 per cent) of the district’s populationof which 84 per cent were males (PACI, 1995). In the absence of a listof individuals residing in the district, we were required to select respondentsin a purposive manner. A relatively comprehensive sample covered the wholedistrict, as well as persons living in various types of housing units, includingfamilies living in rented flats, workers living in company buildings,and bachelors living together in rented flats. Key informants, such assmall shopkeepers or pharmacists in the area, were used to locate therespondents.

The questionnaire was translated into five languages of the sending countries(Bangali, Hindi, Malayalam, Urdu, and Sinhalese). Interviews were conductedby men who spoke the same language as the respondent, and in order to coverthe networking behaviour of men from different social strata within eachcountry, 100 skilled workers and 100 semi-skilled and unskilled workers wereinterviewed. Persons in these occupational categories comprised almost95 per cent of all migrant workers from each country in 1991/92; only 5 per centor less were in professional occupations (Shah, 1997).

366 Shah and Menon

RESULTS

As noted above, about 34 per cent of all respondents had moved throughfriends or relatives, 50 per cent had moved through recruitment agents,15 per cent had arranged their visa directly through the employer, and only1 per cent had arranged their employment through the government recruitmentagency. The latter two groups were added together for the analysis reportedin this article (Table 1, page 377). All the visas by friends/relatives arearranged through a Kuwaiti sponsor (Kafeel), as elaborated in the finalsection of the article.

The major socio-demographic characteristics of respondents who movedthrough friends compared with those who moved through other channels are asfollows. Fifty-eight per cent of those who moved through friends were fromrural areas compared with those who moved directly through the employer(38 per cent), or through the recruitment agent (50 per cent). Compared withthose who moved through recruitment agents, those who moved throughfriends or relatives were older by about two years, and about 71 per cent of bothgroups were married at time of interview. However, only about 2-4 per cent hadtheir family in Kuwait. Unlike those who came through the recruitment agentor friends, almost 20 per cent of those who came on a direct visa wereaccompanied by their family. Seventy-seven per cent of those who camethrough friends were Muslim compared with 54 per cent who came throughagents and only 33 per cent who came on a direct visa. Pakistan and Bangladeshare predominantly Muslim countries, India mainly Hindu and Sri Lankaprimarily Buddhist.

Respondents who came through friends had almost one grade lower educa-tion than those who came through recruitment agents (7.5 and 8.4 grades,respectively). Also, a smaller percentage of those who came through friendshad technical education compared with those who came through the recruitmentagent (21 per cent and 36 per cent, respectively). However, the percentagewho had technical education was highest among those who came by directsponsorship of the employer (48 per cent). The percentage of skilled workerswas also highest among the latter group. In terms of the duration of stay inKuwait, those who came directly through the employer had the longest durationof stay (9.9 years), compared with 6.4 years for those who came throughfriends, and 4.2 years for those who came through recruitment agents. It isanomalous to find that despite their relatively higher education level, those whocame through recruitment agents were earning substantially less than thosewho came through friends (KD 69 and 114, respectively). There are clearlyimportant socio-demographic differences between those who move throughfriends and relatives compared with those who move through the recruitmentagent or on direct sponsorship by employer.


How do friends help?

The most important way in which friends/relatives help in the expansion oftheir social network is by arranging the migration of additional workers.Migrants who move through their friends help in arranging sponsorship forothers. Table 2 (page 377) indicates notable differences in the operation of thenetwork according to nationality of the migrant. For example, 56 per cent ofPakistanis and almost 39 per cent Bangladeshis, compared with only13 per cent of Sri Lankans, came to Kuwait on a visa arranged by friends. Thepercentage who had arranged the visa for another migrant since they came toKuwait was highest among the Pakistanis (36 per cent) and Indians (27 per cent),but only 11 per cent of the Sri Lankans had arranged a visa. Migrants fromvarious countries also differed in terms of access to the social network uponarrival. About 70 per cent of the respondents from each country (exceptSri Lanka, 20 per cent) reported that they had a friend in Kuwait before arrival.Friends provided information for respondents from each country (exceptSri Lanka), 53-70 per cent having heard about their job from a friend in the threecountries.

In case of respondents who reported having friends in Kuwait before arrivingthere, the average number of friends was greatest among the Bangladeshis (5.4)and Pakistanis (4.7). The various types of help received by the respondents isshown in Table 2. Friends provided information, gave financial help, helpedfind a job and in several cases provided accommodation. In terms of type ofhelp received, the Sri Lankans were no different from the others, even thoughthe number who had received such help was much smaller. Networks of socialsupport thus appear to be fairly active in case of most migrants. Despite thepresence of friends in Kuwait prior to their move, not all the respondents whohad friends here arranged their visa through these friends. It is thereforeimportant to explore which factors are most important in determining a movethrough the network of friends and relatives.

LOGISTIC REGRESSION ANALYSIS OF MIGRATINGTHROUGH FRIENDS AND SPONSORING FRIENDS

Expected impacts of the various predictors on facilitating migration

It was expected that older persons would have a greater access to socialnetworks and would be more likely to arrange their move through suchnetworks. We also expected that those from rural areas would play a moreactive role in expansion of the network, since it is usually assumed that the ruralsocial structure places greater demands on helping one’s neighbours andrelatives. Being Muslim was expected to be a facilitator in arranging migration

368 Shah and Menon

for others since Kuwait is a Muslim country. Married persons were expected tobe more likely to sponsor additional migrants since they would be interested inbringing their family to Kuwait. Based on the bivariate analysis, it wasexpected that Pakistanis would be more active in expanding the network. Interms of the education and skill level of migrants, it was expected that thosewith higher educational levels would be more successful in arranging jobs forsubsequent migrants due to their larger network connections. It was alsoexpected that workers who had been in Kuwait for several years, or earnedrelatively higher salary, would have greater access to employers and thereforebe more successful in arranging the migration of workers.

Predictors of moving through friends

In the logistic regression analysis shown in Table 3 (page 378) a move throughfriends is defined as 1 while a move through any other channel defined as 0. Thepredictors in the analysis included the respondent’s demographic characteristics(age, marital status, rural/urban background, religion, and nationality) and hiseducational and skill level (general education, technical education, andpossession of skills). The eight predictors were defined either as binary orcategorical variables.

The eight variables included in Table 3 explained 18.2 per cent of the variancein moving through friends or other channels. Among the five demographicvariables, neither age nor rural/urban background was a significant predictor ofmoving through friends, contrary to expectation. Muslims were more likelythan others to move through friends, and married persons were less likely to doso. Nationality was an especially strong predictor. Compared with Sri Lankans,those from Pakistan were 3.3 times more likely and Indians were 2.3 times morelikely to move through friends. Those who possessed any skills were 1.7 timesmore likely to move through friends than the unskilled. On the other hand, thosewith higher levels of education were significantly less likely to move throughfriends. Possession of technical education did not make a significant differencein the channel through which the respondent moved.

Thus, the characteristics most likely to predict a move through friends andrelatives were being Pakistani or Indian, being a Muslim, and possessing someskills. These findings suggest that Pakistanis and Indians were more successful inusing personal networks for migration. Those with some skills were also able toarrange migration through friends, probably because in their case jobs wereeasier to find. The reason why more Muslims than non-Muslims moved throughfriends could be that jobs for Muslims were perhaps easier to arrange inKuwait.

Being married, and having relatively higher education, were found to signifi-cantly reduce the likelihood of moving through friends. It is possible that


married men, with family responsibilities, would not take the chance to movethrough friends and arrive without a pre-arranged job in the host country. Amuch larger percentage of those who moved through friends arrived in Kuwaitwithout a job compared with those who moved through recruitment agents(32 per cent and 3 per cent, respectively) (Shah, 1998a). Those with higherlevels of education were less likely to move through friends, probably becausethey found it easier to locate jobs directly through the employer and did nothave to depend on informal networks.

Who sponsors additional migrants to expand the network?

About one-quarter of all migrants had arranged the visa of another migrant afterarriving in Kuwait. The results of multivariate analysis to assess the predictorsof such sponsorship are shown in Table 4 (page 379). In addition to the eightsocio-demographic and education/skill characteristics included in Table 3, therespondent’s salary and duration of stay were also included in the analysis ofsponsoring additional migrants. The migrant worker’s salary, as well as durationof stay in the host country, are likely to be important factors in determining hisability to sponsor his family or any other person. Arranging the visa for asubsequent migrant may involve a substantial amount of money, and thoseearning below a certain wage may not be in a position to incur such expenditure.

The ten explanatory variables included in the analysis explain 49 per cent of thevariance in arranging the visa for a subsequent migrant (Table 4). Therespondent’s duration of stay in Kuwait appeared to be an exceptionallyimportant predictor of sponsoring other migrants. Those who had lived inKuwait for nine or more years were 69 times more likely to have arranged thevisa for a subsequent migrant than those who had been in Kuwait for less thanthree years. Also, those who earned a monthly salary of KD 100 (US$330) ormore were several times more likely to have arranged a visa than those whoearned less than KD 50 (< US$165) per month.

Being married and a Muslim were both positively associated with sponsoringan additional migrant. In terms of nationality, we were surprised that once othervariables were controlled, being Pakistani was negatively associated withsponsoring an additional migrant. Further analysis (not shown in this article)indicated that if the respondent’s duration of residence in Kuwait and hissalary were excluded, all three nationality groups (Indians, Pakistanis andBangladeshis) were significantly more likely to arrange additional migrationcompared with Sri Lankans. Thus duration of residence and salary, not nation-ality, were the critical variables that facilitated additional migration. Neitherthe respondent’s age nor belonging to a rural or urban area had a significanteffect on sponsoring an additional migrant. Finally, in terms of the respondent’seducational level and skills, we found no significant effect on arranging the visafor a subsequent migrant.

370 Shah and Menon

The multiplier: how many migrants did the respondent sponsor?

Each respondent was asked the number of visas he had arranged since he cameto Kuwait, and also the number he had arranged after the liberation of Kuwaitfrom Iraqi occupation in February 1991. Seventy-six per cent had arrangedvisas, and about 10 per cent had arranged three or more (Table 5, page 380). Theaverage number arranged was 0.78. Restricting the period to the five or so yearsprior to survey (i.e., since February 1991), each respondent reported havingarranged about 0.42 visas. In terms of the total number of visas arranged by thosewho had arranged at least one visa, the numbers are naturally much larger. Thegroup as a whole had arranged on average about 3.3 visas, Pakistanis 4.5.

When considering sponsorship according to nationality, Pakistanis arrangedthe largest number of visas during their total stay as well as after the liberationof Kuwait: 1.6 visas with only 0.14 by Sri Lankans. After liberation,the difference between the number of visas arranged by the various groups(except Sri Lanka), was relatively small: Pakistanis 0.60; Bangladeshis 0.52,Indians 0.42; and Sri Lankans 0.14. Duration of stay is a very significantvariable in arranging visas for subsequent migrants. Pakistanis had been inKuwait for an average 8.8 years compared with 6.6 years by Indians, 4.5 byBangladeshis and only 3.5 years by Sri Lankans (Table 4). Once the duration ofstay and other socio-demographic characteristics of the respondents were heldconstant, however, Pakistani nationality (of the respondent) was not a positivepredictor of arranging additional migration.

DISCUSSION AND CONCLUSIONS

When a migrant worker is looking for employment in Kuwait, he may arrangehis visa directly through the company or institution hiring him, or use theservices of a recruitment agent to find a job and arrange a visa. Alternatively,he may rely on friends and relatives to arrange a visa and help find a job.Our survey of skilled and unskilled males from four South Asian countries(Bangladesh, India, Pakistan, and Sri Lanka) found that friends and relativeshad arranged the visa for 34 per cent of our respondents. Also, almost one-quarter of the respondents had arranged a subsequent visa for a friend orrelative after arriving in Kuwait. The average number of visas arranged perrespondent was 0.78 for total duration of respondent’s stay; during the five anda half years prior to the survey the average number of visas arranged perrespondent was 0.42. Those who had arranged at least one visa during theirstay in Kuwait had arranged an average 3.3 visas. When considered for thegroup as a whole, the multiplier of sponsorship by friends and relatives appearsto be fairly small. However, even a small multiplier can result in potentiallylarge numbers. Assuming that our findings relate to all migrants in Kuwait, aninitial batch of 100,000 migrants could have arranged the migration of about


42,000 friends and relatives within about five years after the liberation ofKuwait from Iraqi occupation.

The network seems to be playing a more active role in some countries than inothers. For Pakistanis, the visa was arranged by friends in 56 per cent of casescompared with only 13 per cent for Sri Lankans. The average number of visasarranged by Pakistanis was also much higher than for other countries. Thesalience of migrant’s nationality remained statistically significant inthe multivariate analysis and the likelihood of Pakistanis and Indians to movethrough friends remained much higher than for Sri Lankans. However,Pakistanis were no more likely than other nationals to sponsor additionalmigrants once other factors such as respondent’s salary and duration of stay inKuwait were controlled.

The salary of the respondent was a significant predictor that facilitated expansionof the network. Only those earning above a certain ceiling are allowed to bringtheir family with them. Some workers find ways to bring in the family evenwhen they are earning lower wages. However, a certain income level isnecessary to maintain a family, and persons at relatively higher income levelsare therefore more likely to sponsor additional migrants. Fourteen per cent ofthose earning KD 150 or more had their family with them in Kuwait comparedwith only 2 per cent of those earning less than KD 100 per month. Sponsoringan additional migrant may also involve certain costs.

Duration of stay in Kuwait was found to be a highly significant predictorfacilitating expansion of the migrant network, as was found in case of Turkishmigrants to West Germany (Waldorf et al., 1990). It seems that the respondent’sability to sponsor a migrant is enhanced substantially as he spends a few yearsin Kuwait, learns the language and learns about various job possibilities for hisfriends and relatives back home. He is likely to develop contacts with localemployers in addition to his own, and if he has proven to be a good worker heis in a good position to recommend additional workers to his employer. He istherefore likely to become proficient in utilising opportunities in the local jobmarket in Kuwait and also likely to learn about and utilise certain illegal meansof bringing in additional migrants.

The typical arrangement is to buy the visa from a Kuwaiti sponsor (kafeel).Non-nationals may be employed in Kuwait only if sponsored by some institu-tion, company or individual. Since non-nationals cannot own property orbusiness, almost all kafeels are, by definition, Kuwaiti nationals. In order tostart a business, a Kuwaiti national must obtain a license from the authoritiesand obtain permission to import the permissible number of migrant workers tofill the jobs. In most cases, Kuwaitis are unavailable or unwilling to take upmost jobs in the private sector. In 1996, only about 9 per cent of all Kuwaitimales in the labour force were employed in the private sector (PACI, 1996).

372 Shah and Menon

Where a Kuwaiti national does not actually start a business, or is able to obtainmore licenses than required by his business, he can sell the extra visas, thusallowing prospective workers into the country on his sponsorship (kafalat).Although the practice of selling and buying visas is illegal, it appears to befairly prevalent (Kuwait Times, March 27, 1997). A worker must be employedonly by the kafeel who brought him into the country; working for anotheremployer makes him subject to deportation. Since 1996, all of the six Gulfcountries have declared periods of amnesty to allow those in an illegal visastatus to depart without being fined or jailed. More than 11,000 persons leftKuwait during 1997 as a result of the amnesty. It is highly likely that several ofthese persons had arranged their visa through friends or relatives.

The amount of money paid to the kafeel varies according to occupation andnationality of the worker from about KD 100 to KD 250 per year. A work visais bought usually for two years after which it must be renewed by payingadditional money. At the same time, a Kuwaiti kafeel may sell the “permission”for conducting a business on his/her behalf to a non-Kuwaiti. The premises onwhich the business is undertaken are rented out by the Kuwaiti kafeel to thenon-Kuwaiti. Thus, there are several financial arrangements that providesubstantial monetary benefits to the Kuwaiti kafeel and help sustain, andperhaps even encourage, the expansion of social networks. Because arranginga visa requires a substantial amount of money, the salary of the migrant whoarranges a friend’s visa may act as a facilitator in enabling him to purchase avisa. Such a positive association was confirmed in our study. In some cases themigrant worker may arrange a visa for his friend or relative free of charge. Wefound that of those who came through friends, 37 per cent had paid no moneyfor the visa. Considering those respondents who had paid money for the visa,we found that those who moved through friends/relatives had paid a largeramount on average than those who moved through agents, US$4,858 andUS$3,353, respectively.

Those with some skills were more successful than the unskilled in arranging avisa through their friends. Once in Kuwait, however, a person’s skill level wasnot an important predictor of arranging additional migration of friends orrelatives. Those with higher levels of education were less likely to movethrough friends, probably because the chances of such persons finding jobsdirectly through the employer were greater.

Being Muslim appeared as a significant predictor of expansion of the network.Muslims were more likely than non-Muslims to come to Kuwait throughfriends, and they were more likely to arrange additional visas for friends andrelatives. It is possible that the network pressure for sponsoring kinsmen andfriends is greater in Muslim than in non-Muslim countries. Jobs may also bemore easily arranged for Muslims than non-Muslims. In some cases, especiallywithin the Ministry of Religious Affairs, Muslims are required to fulfil certain


positions in mosques or related institutions. Furthermore, some employers mayhave a preference for hiring Muslims, even though general observation does notindicate such a preference in the job market.

Finally, variables such as age of the respondent, rural or urban background, orpossession of technical education did not have any impact on aiding orrestricting the expansion of the social network.

The network of friends and relatives is an effective means of encouragingadditional migration. Whether some of this migration might not have occurredin the absence of networks, or might have been channelled through othersources such as recruiting agents, is difficult of answer. In general, ourobservation and data suggest that previous descriptions of migration throughthe network as a self-perpetuating, self-sustaining phenomenon can also beused to describe the experience of Kuwait. Any estimation of the continuationof such migration in the face of declining job opportunities will need additionalresearch. Some migration through friends and relatives is likely to be high riskin terms of being illegal and such migration is very difficult to regulate andcontrol. However, it is essential that sending governments recognize that asubstantial percentage of all labour migration may now be occurring throughinformal channels. This has implications both for the number of migrantscovered by government statistics and for the long term planning of labourmigration that most South Asian governments are attempting to undertake.

NOTE

1. The research on which this paper is based was funded by the InternationalOrganization for Migration, Geneva, and the United Nations Population Fund,New York.

REFERENCES

Arnold, F., et al.1989 “Estimating the immigration multiplier: an analysis of recent Korean and

Filipino immigration to the United States”, International Migration Review,23(4): 787-812.

Azam, F.1998 “International migration dynamics in high and low migration districts of

Pakistan”, in Reginald Appleyard (Ed.), Emigration Dynamics in DevelopingCountries, Volume II, South Asia, Ashgate Publishers, Aldershot, UK.

Banerjee, B.1983 “Social networks in the migration process: empirical evidence on chain

migration in India”, Journal of Developing Areas, 17(2): 185-196.

374 Shah and Menon

Bocker, A.1994 “Chain migration over legally closed borders: settled immigrants as

bridgeheads and gatekeepers”, Netherlands Journal of Social Science,30(2): 87-106.

Boyd, M.1989 “Family and personal networks in international migration: recent

development and new agendas”, International Migration Review,23(3): 638-670.

Caces, F.1987 Immigration Recruitment Into the Labor Force: Social networks Among

Filipinos in Hawaii, East-West Population Institute, (Reprints of theEast-West Population Institute, No. 209), Honolulu, Hawaii.

Fuller, T.D., et al.1990 “Urban ties of rural Thais”, International Migration Review, 24(3): 534-562.

Gilani, I., et al.1981 Labor Migration from Pakistan to the Middle East and its Impact on the

Domestic Economy, two parts, Research Report Series, Nos. 126 and 127,Pakistan Institute of Development Economics, Islamabad.

Goering, J.M.1989 “The ‘explosiveness’ of chain migration: research and policy issues. Intro-

duction and overview”, International Migration Review, 23(4): 787-812.Gunatilleke, G.

1991 “Sri Lanka”, in G. Gunatilleke (Ed.), Migration to the Arab World,Experience of Returning Migrants, The United Nations University,Tokyo, Japan.

1994 The Economic, Demographic, Sociocultural and Political Setting forEmigration. Sri Lanka Country Paper, prepared for the Mid ProjectMeeting of the IOM/UNFPA Project on Emigration Dynamics in DevelopingCountries, Bellagio, October 3-7.

1998 “The role of networks and community structures in international migrationfrom Sri Lanka”, in Reginald Appleyard (Ed.), Emigration Dynamics inDeveloping Countries, Volume II, South Asia, Ashgate Publishers,Aldershot, UK.

Hugo, G.J.1981 “Village-community ties, village norms and ethnic and social networks:

review of evidence from Third World”, in G.F. De Jong and R.W. Gardener(Eds), Migration Decision Making: Multidisciplinary Approaches toMicrolevel Studies in Developed and Developing Countries, Pergamon,New York.

1994 “Migration as a survival strategy: the family dimension of migration”,in Proceedings of the UN Expert Group Meeting on Population Distributionand Migration, Santa Cruz, Bolivia, 18-22 January, (ESA/P/WP.126).

1995 “International labor migration and the family. Some observations fromIndonesia”, Asian and Pacific Migration Journal, 4(2-3): 273-301.

Jasso, G., and M.R. Rosenweig1986 “Family reunification and the immigration multiplier: US Immigration

Law, origin-country conditions, and the reproduction of immigrants”,Demography, 23(3): 291-311.


Khan, M.F.1991 “Migrant workers to the Arab World: the experience of Pakistan”, in

G. Gunatilleke (Ed.), Migration to the Arab World, Experience of ReturningMigrants, The United Nations University, Tokyo, Japan.

Lindquist, A.B.1993 “Migration networks: a case study in the Philippines”, Asian and Pacific

Migration Journal, 2(1): 75-103.Mahmood, R.A.

1991 “Bangladesh return migrants from the Middle East: process, achievement, andadjustment”, in G. Gunatilleke (Ed.), Migration to the Arab World, Experi-ence of Returning Migrants, The United Nations University, Tokyo, Japan.

Massey, D.S. 1986 “The social organization of Mexican migration to the United States”,

The Annals of the American Academy of Political and Social Science,487: 102-113.

Massey, D.S., and E.F. Gracia 1987 “The social process of international migration”, Science, 237: 733-738.

Menon, R.1988 “How Malaysian migrants pre- arrange employment”, Sociology and Social

Research, 73(4): 257-259.Mullan, B.P.

1989 “The impact of social networks on the occupational status of migrants”,International Migration, 27(1): 69-86.

Nair, P.R.G.1991 “Asian migration to the Arab world: Kerala (India)”, in G. Gunatilleke

(Ed.), Migration to the Arab World, Experience of Returning Migrants, TheUnited Nations University, Tokyo, Japan.

1998 “Dynamics of emigration from Kerala: factors, trends, patterns andpolicies”, in Reginald Appleyard (Ed.), Emigration Dynamics in Develop-ing Countries, Volume II, South Asia, Ashgate Publishers, Aldershot, UK.

PACI (Public Authority for Civil Information)1995 Directory of Civil Information, Population and Labour Force.1996 Directory of Civil Information, Population and Labour Force.

Pohjola, A.1991 “Social networks – help or hindrance to the migrant?”, International

Migration, 29(3): 435-444.Shah, N.M.

1994 “Arab labor migration: a review of trends and issues”, InternationalMigration, 32(1): 3-28.

1997 “Emigration dynamics in South Asia: major findings and policy recommen-dations”, discussion paper submitted for the 12th IOM Seminar on ManagingInternational Migration in Developing Countries, Geneva, 28 and 29 April.

1998a “The role of social networks among South Asian male migrants toKuwait”, in Reginald Appleyard (Ed.), Emigration Dynamics in DevelopingCountries, Volume II, South Asia, Ashgate Publishers, Aldershot, UK.

1998b “Relative success in the host country, Kuwait: does channel of migrationmatter?”, paper prepared for presentation at the Annual Meeting of thePopulation Association of America, Chicago, April 2-4.

376 Shah and Menon

Sighanetra-Renard, A.1992 “The mobilization of labor migrants in Thailand: personal links and

facilitating networks”, in M.M. Kritz, L.L. Lim and H. Zlotnik (Eds),International Migration Systems: A Global Approach, Clarendon Press.

Waldorf, B.S., et al.1990 “A behavioural model of international labour and non-labor migration: the

case of Turkish movements to West Germany, 1969-1986”, Environmentand Planning, 22: 961-973.


TABLE 1

SOCIO-DEMOGRAPHIC AND ECONOMIC CHARACTERISTICSOF THOSE WHO MOVED THROUGH FRIENDS IN COMPARISON

WITH OTHER CHANNELS OF MIGRATION

Characteristics Friends/relatives

Recruitingagent

Direct/government

Total

(N) 275 396 129 800Average Age 33.6 31.7 37.9 33.8Married % 70.9 71.2 90.7 74.3Family in Kuwait % 3.5 1.7 19.6 5.9From rural areas % 57.8 49.5 38.0 50.5Muslims % 76.7 53.8 32.6 58.3Average educational grades 7.5 8.4 9.9 8.3Have technical education 21.1 35.6 48.1 32.6Skilled % 52.0 44.4 62.8 50.0Duration of stay in Kuwait(total mean number of years) 6.4 4.2 9.9 5.9Average salary at present (KD) 114.1 69.4 161.9 99.4

TABLE 2

WAYS IN WHICH FRIENDS/RELATIVES HELPED IN MIGRATION, BY COUNTRY(Percentages)

Type of facilitation Bangladesh India Pakistan Sri Lanka Total

(N) 200 200 200 200 800Respondent's visa wasarranged by a friend 38.5 30.5 55.5 13.0 34.4Respondent arrangedvisa for someone 22.0 26.5 35.5 10.5 23.6Had friends in Kuwait beforecoming 68.5 66.5 71.0 20.0 56.5Heard of job from friends 53.0 69.5 64.0 17.0 50.9

In case of those wherefriends helped, type of helpreceived:(N) 137.0 133.0 142.0 40.0 452.0Average number of friends inKuwait before arrival 5.4 3.1 4.7 1.7 4.1Helped find a job 45.3 83.5 60.6 75.0 63.9Provided information 48.2 85.7 59.9 92.5 66.8Gave financial help 45.3 75.2 59.9 70.0 60.8Arranged visa 57.7 88.7 83.1 70.0 75.8Provided accommodation 43.8 80.5 78.9 82.5 69.0

378 Shah and Menon

TABLE 3

LOGISTIC REGRESSION OF PREDICTORSOF MIGRATION THROUGH FRIENDS

(if came through friends = 1)

Coefficient SE p. value Odds ratio

Age

<30* 1.00

30-39 0.20 0.23 0.399 1.22

40* 0.03 0.29 0.907 1.03

Married status (married = 1) -0.58 0.24 0.015 0.56

Area of residence (urban = 1) -0.03 0.18 0.883 0.97

Religion (Muslim = 1) 0.54 0.26 0.037 1.72

Nationality

Sri Lanka* 1.00

Bangladesh 0.52 0.34 0.129 1.69

India 0.81 0.28 0.004 2.25

Pakistan 1.20 0.36 0.001 3.31

Education

<7 grades* 1.00

8-10 grades -0.57 0.22 0.009 0.57

11+ grades -0.58 0.27 0.034 0.56

Technical education (yes = 1) -0.42 0.22 0.059 0.66

Skill level (skilled = 1) 0.52 0.19 0.006 1.68

Constant -1.07 0.37 0.004

R2 x 100 18.2 %

*Reference category.


TABLE 4

LOGISTIC REGRESSION OF PREDICTORSOF ARRANGING A VISA FOR A FRIEND/RELATIVE

(if sponsored a migrant = 1)

Coefficient SE p. value odds ratio

Age

<30* 1.00

30-39 0.14 0.38 0.716 1.15

40* 0.20 0.43 0.641 1.22

Married status (married = 1) 1.04 0.41 0.010 2.84

Residence (urban = 1) 0.32 0.24 0.186 1.37

Religion (Muslim = 1) 0.97 0.39 0.014 2.64

Nationality

Sri Lanka* 1.00

Bangladesh -0.17 0.53 0.744 0.84

India -0.53 0.41 0.201 0.59

Pakistan -1.16 0.55 0.035 0.31

Education

<7 grades* 1.00

8-10 grades 0.04 0.30 0.902 1.04

11+ grades -0.15 0.38 0.690 0.86

Technical education (yes = 1) 0.27 0.29 0.358 1.31

Skill level (skilled = 1) 0.16 0.27 0.542 1.18

Salary

<50* 1.00

50-99 0.37 0.51 0.460 1.45

100-149 1.49 0.52 0.004 4.43

150+ 1.92 0.55 0.000 6.80

Duration of stay in Kuwait (years)

0-2* 1.00

3-8 2.61 0.75 0.000 13.65

9+ 4.23 0.77 0.000 69.01

Constant -6.66 0.96 0.000

R2 x 100 49.0 %

*Reference category.

380 Shah and Menon

TABLE 5

NUMBER OF VISAS ARRANGED BY THE MIGRANTFOR HIS FRIENDS/RELATIVES

Number of visas arranged(per cent)

Bangladesh India Pakistan Sri Lanka Total

Total 100.0 100.0 100.0 100.0 100.0

0 78.0 73.5 64.5 89.5 76.4

1-2 12.5 18.0 12.0 10.5 13.3

3+ 9.5 8.5 23.5 0.0 10.4

x number of visasarranged* 0.69 0.66 1.61 0.14 0.78

x number of visasarranged after liberation 0.52 0.42 0.60 0.14 0.42

x number of visasarranged by those whoarranged at least one visa 3.14 2.51 4.54 1.38 3.29

*Two respondents said that they had arranged 21 and 50 visas each. The two responseswere recoded as 16 for the calculation of the average. Apart from these two respondents thenumber of visas arranged ranged from 0-16.


LA MIGRATION EN CHAÎNE À TRAVERS LE RÉSEAU SOCIAL:L’EXPÉRIENCE DE LA MIGRATION DE MAIN-D’OEUVRE

AU KOWEIT

La migration de main d’oeuvre vers les pays du Golfe repose principalementsur l’obtention de contrats de travail. Par ailleurs, la majorité des ouvriers dis-posent d’un salaire insuffisant pour aider d’autres membres de la famille àpartir. Malgré cela, les réseaux sociaux se sont développés au Koweit, et leuraction principale consiste en une aide financière, fournie par les travailleursdéjà installés dans le pays, dont bénéficieront d’autres travailleurs migrants.

Cet article a pour objectif de décrire les mécanismes de fonctionnement de cetteaide financière, de présenter les éléments permettant d’anticiper un prochaindépart par l’intermédiaire d’un ami ou d’un parent ou l’octroi d’une aidefinancière à un autre travailleur migrant, et d’évaluer l’effet “multiplicateur” duprocessus ci-dessus. L’article se base sur une étude effectuée parmi 800 migrantsqualifiés et non qualifiés, de sexe masculin, originaires d’Asie du Sud, qui serépartissent comme suit: Bangladesh (200), Inde (200), Pakistan (200) etSri Lanka (200).

Environ 34 pour cent de l’ensemble des personnes interrogées ont pu partirgrâce à l’aide d’amis ou de parents, 50 pour cent par l’entremise d’agentsrecruteurs, et 16 pour cent par recrutement direct ou par le biais des services dugouvernement. Une analyse minutieuse a montré que les éléments déterminantspermettant d’anticiper un départ prochain par l’intermédiaire du réseau socialétaient: être Pakistanais ou Indien, de religion musulmane et posséder unminimum de qualifications.

25 pour cent de l’ensemble des migrants ont facilté l’obtention du visa pour unautre migrant depuis leur arrivée au Koweit. Cette même analyse a montré quela durée du séjour au Koweit de la personne interrogée constituait une indica-tion particulièrement significative quant à la probabilité d’une aide financièreen faveur d’un autre migrant.

Un salaire mensuel, le fait d’être marié et de religion musulmane sontégalement considérés comme autant d’éléments déterminant liés à une aidefinancière future. Concernant l’effet multiplicateur, on peut affirmer quechaque migrant a aidé à l’obtention d’une moyenne de 0,78 visa. LesPakistanais, dont la durée moyenne de séjour au Koweit est la plus longue, ontcontribué à l’obtention du plus grand nombre de visas (1,6 en moyenne).

382 Shah and Menon

CADENA DE MIGRACIONES MEDIANTE REDES SOCIALES:LA EXPERIENCIA DE TRABAJADORES MIGRANTES

EN KUWAIT

La migración de trabajadores a los países del Golfo es sobre todo de tipocontractual y la mayoría de los trabajadores no llegan al nivel salarial necesariopara poder acoger a miembros de la familia. Pese a ello, en Kuwait se han idoextendiendo redes sociales mediante las cuales los que ya se hallan en el paíspatrocinan a nuevos trabajadores migrantes.

Los objetivos del presente artículo consisten en describir de qué manera actúael proceso de patrocinio, determinar qué elementos permiten predecir movi-mientos a través de amigos o parientes, o disponer el patrocinio de nuevostrabajadores migrantes, y evaluar el efecto “multiplicador” del citado proceso.El artículo se basa en una encuesta realizada con 800 migrantes de Asiameridional, de sexo masculino, cualificados y no cualificados. Procedían deBangladesh, India, Pakistán y Sri Lanka, 200 de cada uno de esos países.

El 34 por ciento aproximadamente de todos los que respondieron se habíandesplazado mediante la ayuda de amigos o parientes, el 50 por ciento medianteagentes y el 16 por ciento por contratación directa o a través de una oficinagubernamental. El análisis de regresión logística indicaba que los factores quemás probablemente permiten predecir un desplazamiento mediante la redsocial eran ser paquistaní o indio, ser musulmán y poseer alguna cualificación.

Desde su llegada a Kuwait, una cuarta parte de todos los migrantes habíantramitado algún visado en beneficio de otro migrante. El análisis de regresiónlogística mostró que la duración de la estancia en Kuwait del corresponsal eraun importantísimo elemento de predicción en cuanto a la tramitación delpatrocinio en favor de otro migrante.

También los ingresos mensuales, el estar casado y el ser musulmán seasociaban positivamente al patrocinio. En cuanto al efecto multiplicador, cadamigrante había tramitado aproximadamente 0,78 visados. Los paquistaníes,que son los que más tiempo han permanecido por término medio en Kuwait,presentaron asimismo el máximo promedio de visados: 1,6.

Chain Migration Through the Social Network: Experience of Labour Migrants in Kuwait

Documents