Because of the conventional approach of defining migration, the intensity of migration is generally reported to be very low in India. The two major sources of data on migration in India, the Census and the National Sample Survey (NSS) cover only permanent or semi-permanent migration with seasonal migration partly overlapping with the category of short duration migration (Srivastava 1999). However, even the definitions of permanent or semi-permanent migration adopted in the Census and the NSS do not consider long term migrants without changing the place of residence as migrant. As a result the coverage of these two data sources becomes so restrictive that both of these report declining trends of out migration over the years (Kundu and Gupta, 1996). Contrary to this The National Commission on Rural Labour (NCRL, 1991) finds increasing trends of seasonal and temporary migrants over the years. Besides this, many of the micro studies dealing with the issue of migration in the theoretical context as well report much higher incidence of out-migration, particularly in the case of underdeveloped regions (Breman, 1978, 1994; Ghosh, and Sharma 1995; Bora, 1996; de Haan, 1999). Besides, there is a considerable change in the nature and cause of migration---in recent years proportionately more migration is taking place among the labour force in search of their livelihoods, which is mainly for relatively longer terms (de Haan, 2000; Rodgers and Rodgers, 2000; Sharma et al., 2000).

The origin of current migration from Bihar can be traced back to the days of ushering in green revolution in the north-west India, which created unprecedented demand for labour in agriculture. During the early 1970s it started with a small trickle to the rural areas of Punjab which soon took the form of a flood during the next one decade. Soon the influx of migrant labour to rural areas of Punjab started spilling over to the neighbouring state of Haryana where the green revolution was making a headway. However, the rising trend of migration to rural Punjab stabilised after early 1990s for various reasons: prolonged period of terrorism and attacks on migrant labourers during this period, changing cropping pattern to suit the market demand that allowed extensive mechanisation of agriculture, resulting into the displacement of rural labour, and also relatively slow rise in wage rates. These together dampened the flow of migrants from rural Bihar to rural Punjab (Singh 1995). Partly the flow was diverted to new areas such as Haryana and even to the parts of Jammu but, on the whole, the number of migrants to Punjab and Haryana either stabilised or started declining.

The restricted demand for migrant labour in the green revolution areas does not mean that over the years the number of out-migrants from rural Bihar has also come down. In fact, during the last two decades the phenomenon of migration from rural Bihar has taken up an alarming proportion. In fact, 'members of the ubderclass' are migrating in large numbers in search of better employment, which is evident on the basis of both village-based inquiries and macro evidence (Brass, 1993; Jha, 1997; Wilson, 1999; Chakravarty, 2001). During the course of time not only the absolute number of out-migrants has increased but also the rate of out-migration has almost doubled over this period.

Against this background the present paper attempts to highlight the emerging trends and pattern of labour migration in rural Bihar. Besides this, one of the general objectives of this study is to examine the socio-economic factors behind labour migration from rural areas of Bihar and its impact on the individual migrants, their families as also on the village society and economy as a whole. The study is based on the primary survey data collected at two points of time after a gap of approximately 18 years from randomly selected 6 villages of north Bihar. Out of these 6 villages 2 each are from three districts viz. Gopalganj, Madhubani and Purnea. The data has been collected under two major research projects conducted during 1981-83 and 1999-2000. Besides this, the later study (1999-2000) also conducted a census survey in 18 villages (including above mentioned 6 villages) covering a population of more than 38 thousands from more than 6400 households. The data collected from census survey has been largely used to evaluate the socio-economic conditions of the study area, and migrants and to examine the intensity of permanent migration from the villages under study.

The paper has been arranged in four sections. After presenting introduction in the first section, the second section (Section II) deals with the socio-economic background of the study area in general and of the migrants and non-migrants in particular. This section is based on the census survey conducted in the 18 villages of the three districts and briefly presents an account of socio-economic factors responsible for out-migration. In the third section (Section III), based on sample households from 6 villages from the same districts, the changing nature and magnitude of migration from rural Bihar has been presented. Section IV deals with some other details of current migration and the likely impact of out-migration on the households. In the last section, i.e. Section V, major conclusions and policy recommendations have been presented.

In this section a brief account of socio-economic characteristics of migrants and non-migrants has been presented. Presentation of these basic socio-economic characteristics for migrants and non-migrants separately helps in understanding the difference in their socio-economic backgrounds. This is also helpful in understanding the answer to the question ‘who migrates?’ However, before going into the question of ‘who migrates’, we will try to identify migrants from the study area. Identification of migrants is difficult because nature of migration from rural areas is quite diverse. People migrate for different time periods and different purposes. Here all those persons have been considered as migrants who left the village at least for more than two months during the last one year of survey. To be more specific following definition for the identification of migrants has been followed in the present study.

1. Migrating persons: All those persons have been considered as migrants who have either migrated for at least two months during the last one year at the time of survey. However, if a person has returned from destination and is not likely to migrate again in the same year he has not been considered as migrant.
2. Short term migrant: All persons migrating for a total of less than nine months during the last one year has been considered as short term migrant.
3. Long term migrant: All persons migrating for a total of nine months and more during the last one year has been considered as short term migrant.
4. Migrating households: On the basis of the definition of migrant mentioned in a), all households reporting migration of at least one of the family member, has been defined as migrating households. Accordingly, all those households who do not report migration of any of the family member have been categorised as ‘non-migrating households’.

After defining migrants on the basis of above definitions, the census survey in the sample villages revealed that there are more than 13 per cent of total population who migrated for at least two months or more during the last one year. There are more than 48 per cent of the households, who reported at least one migrant from their family. Some of the basic socio-economic characteristics of migrating and non-migrating households are presented in Table 1.

The caste groups considered here are largely based on the State list of caste groups. However based on certain distinguished characteristics, some changes in the groupings has been done. Within OBC II, Kurmi, Yadav, and Koeri have clearly exhibited some distinct characteristics and are comparatively more upwardly mobile castes. Hence, these castes have been clubbed under OBC II (upper). Muslims have been although classified under the four categories identified by the state, they have been treated separately in the present study.

The caste composition of the migrating and non-migrating households shows that Muslim households are the most prone to migration followed by OBC I and upper caste households. The representation of Muslim and OBC I in migrating households is more in comparison to their respective representations in total number of households. The representations of SC and upper castes are slightly lower to their respective share in the total number of households. The least prone to migration are the households belonging to OBC II, both OBC II (upper) and Other OBC II. However, the reasons of low migration in these two caste groups are different. In case of OBC II (upper) most of them are cultivators and somehow remain busy with the stamp piece of land they own. Yadav and Kurmi among them are also involved in wide scale leasing in of land, which keep them relatively more tied in the village. Other OBC II group is largely comprised of artisan castes such as lohar (blacksmith), sonar (Goldsmith), kumhar (potter), etc. who keep themselves busy with some kind of non-farm works round the year, and hence are less prone to migrate.

Although share of nuclear family is higher in all migrating households because of wide prevalence of nuclear family in the region, the proportion of migrating households in all households is much higher in case of joint families. In fact, in a joint family it is easier for the households to release some of their family members for migration. Table 1 gives the proportion of migrating households from nuclear and joint types of family structure. The percentage of joint family in all the households is 26 while among migrating households this percentage is more than 34.

The land structure of the households shows that the proneness to migration is almost same among all the land classes. From all the land classes the proportion of migrating households is almost equal to their respective proportions in the total number of households. However, it is evident from the Table 3 that more than two third of the migrating households hail from either landless or very small size (up to 1 acre) of land holding class. Moreover, once again we will see later in the next section that the nature of migration from the lower and the higher land classes are distinctly different.

One of the ways to determine the status of households is the type of main source of livelihood. For this purpose occupation of head of the household has been taken in to consideration. Although the occupation of the head of the household is not the sole occupation, which all the family members depend on, the same may determine the status of the household to a larger extent.

If head of a household is cultivator or engaged in some kind of non-farm petty business, there is little possibility for family members to migrate. This is evident from the fact that the proportion of cultivator households is much lower among migrating households in comparison to that among non-migrating households. The most prone to migration households are those, whose head of the households are either non-agricultural labourer or are involved in some private service.

The asset position of households shows that among the migrating households nearly half of the total number of households possesses very little assets.

Among the households where asset position is better the tendency to migrate is less. It can be seen from the table 5 that approximately 50 per cent of the households possess assets worth Rs. 1500 or less, where intensity of migration is higher. In the lower asset value households the representation among migrating household is higher than that among the total households. It necessarily indicates that poor households are more prone to migrate.

The foregoing discussion on the types of households suggests that although migration is widespread in Bihar, most of the migrants come from lower caste and class of the households. Abject poverty and lack of employment opportunities in villages has forced a large proportion of populace to migrate to earn bread and butter for their family members

Age analysis of migrants in Table 2 shows that the major chunk of migrants, whether for short duration or long duration, hail from the age group of 15 to 39, which is a potentially most productive and creative age in the life span of the individuals.

The detailed analysis of age, in fact, shows that the maximum number of migrants hail from the two age categories, namely, 15-24 years and 25-39 years.

If we analyse data on gender and migration, the effects of migration on gender distortion can clearly be seen. In rural Bihar there are more female children born than male children. By the time they reach 14 years males out number females, their respective percentages are 54.31 and 45.69. It means girl children, though naturally gifted with better survival potential, become victim of neglect due to various social reasons. This distortion in relative gender proportion continues till end. In rural Bihar the problem of gender becomes much more acute with the migration of male who are potentially capable of working elsewhere in India. In the age group of 5-14 only 2.67 per cent migrate. That is why even after out migration male children continue to dominate over female children in number, even though the gender gap is slightly reduced. However, in the age group of 15-59, more than 95 per cent were male migrants. Consequently in the rural areas of Bihar, out of the population left behind in this age group of 15-59, only 40.5 per cent were males and the remaining 59.5 per cent were females. The absence of a large number of male population is not without consequences both for women and children. In the long run it may result into serious socio-economic and psychological crisis in rural Bihar.

For evaluating the changing trends and pattern of migration from Bihar data from sample survey from six villages has been analysed. The change in the trends and pattern of migration has been recorded over a period of approximately 18 years. As has been mentioned earlier, the data has been taken from two surveys conducted in the years 1981-82 and 1999-200 in the same six villages.

Based on these two surveys it has been estimated that during the intervening period the intensity of migration has almost doubled. A glance at Fig 1, gives a clear picture of the change in the trends of migration from rural Bihar.

Fig 1: Percentage of Households Reporting Migration and Persons Migrating during 1982-83 and 1999-2000

Fig. 1 suggests that there were 27.69 per cent households reporting migration in 1982-83. By 1999-2000, there is a steep increase in the number of households with at least one migrating family member (hereby referred to as migrating households) and their percentage jumped to 48.63. It means approximately every alternative household is effected by migration, whether for a short or long duration, depending upon the whole host of circumstances. Further, in 1982-83 the migrant population constituted 7.49 per cent of the total population, which steeply increased to 13.42 per cent by 1999-2000. This stark reality of rural Bihar clearly defies all speculation of declining trends of migration from Bihar. A more detailed comparison of the changes in the trend of migration from the three districts under study can be seen in terms of caste and class composition of migrants (Table 3).

Caste-wise break up of the migrating households shows that over the last 18 years there is increase in the percentage of migrating households across castes but the rate of increase is much higher among the OBC II, followed by SC, and Muslims. As such, OBC I households migrate in highest proportion, that is, 53.12 per cent of the households are sending one or more of their family members out of the village of residence. Though the rate of increase of migration is the highest among the OBC II, but within this group of castes, Yadav, Koeri, and Kurmi are least interested in migration. In fact, these three castes of OBC II send the smallest percentage of migrants out of their village of origin. As far as percentage of persons migrating out of their own respective group of castes is concerned, upper castes are constituting the highest proportion in 1982-83 (10.24% against the overall figure of 7.49%) and they have maintained their highest percentage even during 1999-2000 (15.20% against the overall figure of 13.42%). This only indicates that migration from Bihar is not simply of distressed masses; there is a substantial proportion of migrants going out with the hope of upward mobility. However, the higher rate of increase in the proportion of migrants from the OBC II, SC and Muslim population shows that over the years there has been increasing tendency of migration also among relatively less mobile population groups.

* Social Class: This category has been introduced for making two sets of data (1982-83 and 1999-2000) comparable. The abbreviations used in the Table are: AL= Agriculture Labour; PMP= Poor Middle Peasant; MP=Middle Peasant; BP= Big Peasant; Landlord= Landlord; and NAG= Non Agriculture

Class analysis further brings out clearly the two different types of streams of migrants flowing out of Bihar. One is constituted of agricultural labourers and the other one of landlords and big peasants. Both labouring households and upper classes are trying to desert rural Bihar for their own set of reasons that are not same for both of them. It is clear from the data that in 1999-2000, 53 per cent of the landlord households are contributing at least one member to the stream of migration, and 17 per cent of the people belonging to this class are part of the migration exodus. At the other end of the economic hierarchy, 48 per cent of the agricultural labourers’ households are sparing their members for migration. Out of the total population of agricultural labourers, 11.39 per cent had joined the stream of migration. The land ownership status of the households also gives similar results. Migration is higher from among the lowest and the highest rungs.

In comparison to 1982-83, in the year1999-2000, there is a major shift in the pattern of migration. The propensity to migration in the middle categories of land ownership (2.5 to 5 acres) has come down but the same has shifted to the extreme ends of the land ownership hierarchy. Half of the households who are either landless or owning one acre of land are sending at least one member out of the village of origin. It means among the land starved categories every second household is surviving by migration. Nearly 14-15 per cent of the population, pertaining to above two categories, go out for making two ends meet. The categories owning land between 1-5 acres are trying hard to survive by staying out. However, the intensity of migration increases among the upper echelon of landed hierarchy. More than 60 per cent of the households and 14-15 per cent of their population is out-migrant.

During the early 1980s the general preference of the migrants was the short term migration. More than three-fourth of the migrants was staying away for the shorter duration. However during the next two decades the trend reversed drastically in favour of long term migration (Fig 2).

Though all the social categories during the early 1980s did tend to stay at the place of destination for shorter period, but this tendency was relatively stronger among the lower castes and poor classes (Table 4). Partly it was so because of their rural to rural migration and the seasonal nature of the availability of employment.

By the year 1999-2000, among the upper castes 93 per cent of the migrants are migrating for the longer duration and only a small proportion of 7 per cent are migrating for short duration. The same is true of OBC II group of castes. As we go down along the caste hierarchy, the duration of stay at the place of destination also decreases. This is generally true of class hierarchy. More than half of the migrants hailing from the ranks of agricultural labourers is staying outside their village for short duration. On the other hand, 80-90 per cent of the migrants belonging to middle peasant, big peasant, and landlord class is staying away for longer duration. It is also true of the higher land owning categories of rural Bihar. Migrants from the landless households and from those owning less than one acre of land tend to migrate for shorter duration. Even among them more than two-third of the migrants stay away for longer duration. It only shows that the predominant trend among the current migrants from rural Bihar is towards long term migration, though a small proportion among the lower castes/classes still migrate for short duration for various reasons.

Change in occupational structure of migrant workers, over the years has been depicted in Table 8. Almost half of the total migrants during 1982-83 were non-workers. Out of the remaining half, some of them were working at the urban places while others were attracted to rural areas. Among those migrating to various urban centres, a large proportion was absorbed in the service sector, both private and government. Migrants who were going to rural areas, a large proportion of them were performing agricultural labour. Over the years a shift has been witnessed in the occupational structure. A clear change observed in the occupation status of the migrant workers is increasing casualisation of the workers during the intervening period (Fig. 3).

Fig 3: Changes in Occupational Structure of Migrant Workers

In comparison to 1982-83 there has been sharp increase in the proportion of casual labour in 1999-2000. Besides there has also been increase in the rank of workers involved in private service. Since most of the workers involved in private service are mainly working in small industries, trading, etc. on ad hoc or temporary basis, their conditions are also not much different from daily casual wage earners. The percentage of workers involved in these two categories in 1982-83 were approximately 18 and 10 per cent respectively. This percentage increased to approximately 34 and 31 per cent respectively during 1999-2000. The major contribution to these two categories of workers have been from the ranks of non-workers.

By 1999-2000 the percentage of non-workers among the migrants has come down from almost 50 per cent to 25 per cent (Table 5). Now a larger percentage of workers are able bodied young males who migrate with a clear purpose of earning, and thus, the proportion of non-workers has come down drastically. Those who decide for urban place, major chunk of them are absorbed into the private service sector, followed by casual labour. A small percentage is also absorbed into the government services but they must be privileged migrants only. There are three major occupations for the migrants in the rural areas. The largest proportion is performing agricultural labour, followed by non-agricultural labour and private service.

After the discussion of broad contours of changes in the trends and nature of migration since the early 1980s, we will now concentrate over some other dimensions of current migration.

Search of new and/or better employment has been one of the most important reasons for migration of workers. In fact more than two third of the migrants gave ‘search for employment’ and ‘better wage’ as the primary reason of their migration. However, there are also various other socio-economic reasons, which contribute to large scale exodus of population from the region (Table 6).

Here we have also identified the reasons for migration and, with a particular end in mind, for how long a migrant stays away from the village. Although marriage has been one of the most important reasons of migration, particularly in case of females, this has been kept out of our analysis in order to focus on non-voluntary causes of migration. Some of the reasons for migration identified during the course of data collection, were employment (on hired basis), better wages, self employment, study, avoiding factionalism within the village, accompanying with migrating family, job transfer, etc.

It is not possible to make a neat distinction between the two intimately linked reasons, that is, hired employment and better wages. Better wages do not mean anything if there is no employment available at that wage rate. ‘Sufficient’ employment at better wage rate seems reason if we combine the two. Out of the total migrants from rural Bihar 77 percent migrate with the hope of getting sufficient employment at enhanced wage rate (Table 6).

For the purpose of analysis if we ignore the reason of better wages, and limit only to the employment as prime reason for migration, even then there were 38.7 percent migrants in this category. These are the migrants who more appropriately be called as distressed migrants. They migrate not for better economic opportunity with higher wages at the place of destination, rather they migrate because in their own village there is no employment at all, at whatsoever wage rate. The fear of starvation for want of employment pushes them out of the rural areas, with no clear choice of destination. Those migrants who migrate for better wages, do make choice of destination and occupation to engage into therein. They too live under severe economic constraints at their respective village of origin. The difference of their distress is only of degree and not of kind. Migration in search of self-employment or caused by family ties, what we better call as string-migration, is rather limited. After ‘search for employment’ or better wages, the third most important reason for migration is entry into the job market or transfer of service. In this category there are 13.6 per cent migrants is a substantial number. The above data depicts two different types of streams of migrants, pouring out of rural Bihar, very clearly. The major stream is constituted of distressed migrants. They hail largely from the ranks of rural landless agricultural labourers or are contributed by the poor peasantry. On the other side, the second stream of job seekers seems to emerge out of the households of rural elite who are socially and economically better placed. There is a direct relation between the total duration a migrant would stay away from his village and the purpose with which he migrates. The out-migrants who quit village with the hope of employment or better wages, largely would try to stay away at least for 5-6 months. If they can manage their employment for more than 6 months, they would consider themselves lucky. At the same time, most of the migrants from this stream of mass-migration like to have live link with the place of origin. The reasons may be social or economic. In fact, migrants hailing from the households of poor peasants own stamp size plot of land which they would like to cultivate with the help of their family members. This anchorage of land encourages seasonal visits to their home village and never let them snap their relation with the ‘motherland’. On the other hand this is not the case of white collar migrants who have gone out to join some regular jobs or have migrated due to transfer of service. More than three-fourth of the migrants from this category were those who stayed away this category were those who stayed away for more than nine months.

Migrant workers get information about availability or prospects of availability of work at destination from different sources. There has been enough literature to show that earlier middlemen or contractor used to be an important source of information for the migrant workers. However, in recent times although the role of middlemen has not entirely been replaced by other sources, their role has been minimised largely because of increasing awareness among migrating workers.

Of the total 323 migrant workers from 267 households 41 percent got the information from their friends/relatives and another 35 per cent just accompanied their already migrating co-villagers. It means that for three-fourth of the migrants the source of information is either the co-villagers or the kinsmen. Moreover, it is interesting to note that the role of middleman/contractor has been more prominent in case of short term migration (44.44%), while the role of friends/relatives has been more prominent in case of long term migration.

One of the most import findings of the study is to map out successfully the clear links between the origin and destination. More than one-fourth of the migrants from rural Bihar goes to Delhi in Search of employment. Most of the migrant labourers are absorbed into the occupations such as, rickshaw pulling, building construction, carpentry or masonry work, and various other types of casual work in informal sector.

Table 7: Percentage Distribution of Migrants by State/City with Rural Urban Region of Destination

More often than not they are absorbed at the lower rungs of the segmented labour market. Certainly those labourers would go to Delhi who are relatively smart and wield better bargaining power in the market as they have to live in an hostile urban environment of Delhi. The destination for another 25 percent migrants is the '‘greener pastures'’ of Punjab and Haryana. Of the total migrant workers coming to Punjab and Haryana four-fifth are absorbed into the rural areas and the remaining primarily worked it the urban places of these two states.

There are 6 per cent intra-state migrants who primarily opted for urban places of Bihar. Uttar Pradesh, West Bengal, and the North-East were the next places of destination. Nearly 5 to 6 per cent of them migrated to each of this place and by and large worked at the urban places in petty jobs in secondary and tertiary sector. There are 4.4 percent migrant workers who were attracted by the city of Bombay – a dreamland for the migrants. The most interesting aspect about their choice of destination is the place like Sikkim and Rajasthan, though at both these places they largely confined to one or two urban centres. There is also 1.4 percent such migrants who choose their destination abroad, particularly the Gulf countries. In the beginning Muslims went to the Gulf but now even Hindu population is also opting for such destinations at abroad (Table 7).

Income and earnings of the migrants at the destination depend largely on the types of occupations they are involved into and the duration of the job available. Besides, it also depends on the personal endowments such as level of education, skill, years of experience, etc. of the migrant workers. Since most of the migrant workers possess low personal endowments, they get absorbed generally in informal sector, some kind of irregular or casual employment, such as, casual wage earners, domestic servants, earth digging, private tutoring, etc., their average earnings are abysmally low. In fact the migrant workers face two types of employment and income insecurity at the destination. First, most of the workers do not get employment on regular basis and hence they have to struggle against the insecurity of unemployment. Second, since most of them are involved in lower quality of employment they get remuneration at the subsistence level. The earnings of most of the workers are so low that it becomes difficult for most of the migrant workers to save money and remit to their families. However, declining employment opportunities at the native place and expectations to get more remunerative employment at destination keep the migrant workers tied to the place of destination. The average income size of the migrant workers is presented in Table 8.

There are various factors that effect the average amount of monthly earning of a migrant workers. Here the effect of literacy, age at first migration, total years of migration and occupation on the average amount of monthly earnings has been studied. The overall average amount of monthly earning of a migrant worker comes out to be Rs. 3373/-, which is not a small amount compared to the real or statutory wage rate in rural Bihar. However, there are two problems with the overall figure. One, it includes the earnings of high salaried migrants, which affects the average figure. Two, it is the income pertaining only to the period of employment and do not adjust the periods of unemployment of the migrants during the year.

If these two factors are taken into account the average monthly income of a migrant workers, who is most hard pressed, would be far below. Higher the level of literacy, higher is the average monthly earnings of a migrant. This is truer in case of the long term migrants than in case of the short term migrants. The age at the time of first migration also has bearing upon the average monthly income of the migrants. Higher the age at first migration, higher is the monthly income. While this is true for the long term migrants, it does not hold good in case of short term migrants. In the latter case those migrants are earning highest who migrated at the age between 21-25. Below this age and above this age the average monthly earning of the migrants are low. There can be varied reasons for the reverse relation beyond a specific age limit. Those migrants who join migration stream too early must be economically hard pressed at home and ready to work on anything that could give them relief. This sets the vicious circle of low earning and low spending. These migrants cannot afford gestation period for acquiring skill that can fetch them better income later on in their life. Also a substantial chunk of short term migrants go for agricultural labour who requires least skill, and it is possible that the younger migrants join the agricultural work and stagnate in terms of earnings. Generally 20-25 is the age when a young male migrant aspires to rise in the economic ladder and he is in a better position to take various challenges. Beyond this age the burden of family responsibilities do not leave much space for experimentation, and thus, older age migrants are also pushed to the low income sector of the economy. Occupational classification of the migrants also highlights income disparities. Non-agricultural labourers are the worst paid and those working in the government jobs are the best paid. The reason for low income in non-agricultural sector is irregular employment. To some extent this is true of the migrants absorbed in private service sector. It is for the same reason that the average monthly income of an agricultural labourer comes out to be more than those migrants working in non-agricultural sector or private service sector. However, it is to be remembered over here that even the agricultural labourers are employed only for a few months during the peak periods of labour demand and for the rest of the period they have to fend for themselves. There is a clear relation between the year of first migration and the average amount of monthly earnings. Longer is the experience of migration, higher is the amount of average monthly income. This is largely true of short term as well as long term migrants.

It is important to throw light as to what extent the migration helps to fulfil the dreams of the migrants. It is true that the expectations would vary with socio-economic background of the migrants and many other factors. But still it is not out of order to quantify the economic gains of the migrants.

On the whole the remittances constituted nearly one-third of the average annual income of the households under study. By any standard it is a huge amount even if the overall average income of the households is low. There is not much of the inter-district variation as far as the average amount of the remittances is concerned. However, since the average annual income per household varies in all the three districts under study, accordingly the percentage of contribution due to remittances also shown variation. The caste/class relation of the remittances brings out the details of the difference in the amount of earning by way of migration. It is generally true that higher is the caste, higher is the average amount of the remittances earned.

However, among the SC/ST and Muslim households the contribution made by remittances turns out to as high as 29 per cent. It means, in relative terms, even the small amount of the money earned by migration gives them a big economic relief to their meager income. Similarly, even the small amount of remittances is a big help to the labouring poor in rural Bihar. Among the households of landless agricultural labourers and marginal peasants, remittances may add one-third to their existing average income per household.

It is well established that most of the migrants quit their place of origin in search of better opportunities elsewhere. We have already established that most of the migrants from rural Bihar out-migrate under distressed conditions. It is therefore, important to investigate that to what extent migration has helped them to ameliorate their economic condition. Data on remittances have been collected. Table 10 provides information on the relation between caste, class occupation and the money remitted by the migrants.

Out of the total migrants in the sample study, 49 percent are remitting money regularly, 37 percent occasionally and there are only 14 percent who did not send any money back home. There is no major inter-district variation with respect to remittances.

There is no significant difference between the overall percentage of migrants remitting money and the landless labourers or marginal peasants. Therefore, there is no significant variation, with respect to the proportion of those who remit money (regularly or intermittently) and those who do not. The same is largely true of different castes. Though among the upper castes there were only 9 percent such migrants who did not remit any money the percentage is higher among the lower castes. Among the SC/ST migrants 22 percent does not remit any money back home. The relation between the frequency of those remitting money and occupation of the head of the household is also reflected upon in Table 10. It is found that 27 percent of the migrants belonging to the agricultural labouring households could not remit any money. The above discussion leads to one general conclusion that there seems to be a close relation between deprivation at home and low frequency of those remitting money, whether regularly or intermittently.

Is there any relation between the average amount remitted by a migrant and his/her class, caste or occupation? The overall average amount of money remitted by a migrant is Rs. 8105/-. It is interesting to note that those migrants who do not own any productive assets are remitting money above the overall average of Rs. 8105/- whereas those who own productive assets upto Rs. 5000/- are remitting below the overall average.

It is possible that the abject poverty of the migrants forced them to space larger amount for remittance, even when their wages/remittance, even when their wages/earnings might be abysmally low. This pressure on other poor migrants being relatively low, they were not squeezing themselves to the hilt. However, those migrants owning productive assets worth of more than Rs. 5000/- are able to remit money above the overall average, may be because of their better earnings from their place of destination. Interestingly again same is true of caste status and the amount remitted. Upper castes and SC/ST category of castes are remitting the highest amount of money, while the middle castes and Muslims are remitting, on an average, less than the overall average amount of Rs. 8105/-.

The occupational classification of the migrant households throws light on another important aspect of the average amount of remittances. The highest average amount of money is remitted by those migrants who belong to households doing government service. In fact, these migrants remit more than four times of the overall average amount of remittance. Those migrants in private service also reemit, on an average, a little above the overall average. The remaining all the migrants, on an average, remit money ranging from Rs. 5 to 7.5 thousand annually. Though landless migrants and those owning less than one acre of land remit below the overall average, it is the landless among both of the categories, which remit higher amount. Households owning higher amount of land also seems to be in government service and thus, are in a position to remit higher amount of money.

Our above formulation is further confirmed by the direct correlation between literacy and the proportion of migrants remitting money. Out of the total migrants, 39 percent were illiterate. Among the illiterates 18 percent were those migrants who could not remit any money (Table 11).

Those migrants who are literate are little better off and the percentage of migrants who could not reemit any amount is 14. As the literacy level of migrants go up the proportion of those remitting no money comes down. On the whole, also higher is the level of education of migrants, higher is the average amount of money remitted which is understandable. For good government and private jobs there is a need of minimum educational qualification. That is why educated migrants are able to remit higher amount compared to less literate or illiterates. Marital status of migrants is demographic dimension that certainly affects the average amount remitted. Three-fourth of the migrants is married. They remit not only on more regular basis but also in higher amount. Fifty five percent of the married migrants, on an average, remitted Rs. 9346/- on regular basis. Another 33 percent are not regular. However these proportions are reversed in case of Unmarried migrants. Not that they remit average amount of only Rs. 4695/- there are only 33 percent among them who remit that amount on regular basis. Another 20 percent do not remit any amount; the remaining 47 per cent remit only off and on.

There also seems to be a direct correlation between the year of first migration and the number of those migrants remitting money on regular basis. The early migrants, that is, those who started migrating to other places prior to the decade of 1990s, are not only remitting higher average amount but are also remitting it on relatively regular basis. The same is not true of migrants who began to migrate during the 1990s.

The data from our census survey throws light on how the households, hailing from different districts and belonging to various castes and classes, use the remittances (Table 12).

There are two major heads of expenditure among the migrants: daily consumption needs and medical treatment. Out of the total savings 38 per cent is spent on meeting daily consumption needs and another 27 percent on medical treatment. Though more than four-fifth of the migrants took debt, before departing from Bihar, to meet the travel costs, only 12 percent of the earnings went into the clearing of the debt.

The above pattern of household expenditure is quite in line with the studies made in the past on the household budget of labouring poor. But there is one contrast with the pattern of expenditure among the households of local labourers. There are two major causes of prevalence of large scale indebtedness among the local labourers, that is, expenses on marriage and medical treatment. In the case of migrants the expenditure on marriage ceremony is less than two per cent and on the development of agriculture was also as low as 4 percent. It reinforces our earlier findings that the mainstream migrants from rural Bihar decide to migrate for employment under distressed conditions. After hard work, the meagre savings first go to meet daily consumptive needs, followed by expenditure on health and debt repayment. And this pattern f expenditure is largely true the migrants from all the three districts under study.

Similarly there does not seem any major inter-caste differences with respect to the pattern of expenditure. Across caste the major expenditure is incurred on items of daily consumption followed by medical treatment and consumer durables. However, there is star k inter-caste variation with respect to the use of savings on debt clearance and agricultural development. Upper castes used less than six percent of their savings in debt clearance while in the case of SC/ST category almost one-fourth had gone into the debt clearance.

Reverse was true as far as investment on debt clearance is concerned. Against the overall is per cent investment in agricultural development, OBC II (Upper) invested more than 10 percent of their savings while SC/ST invested not even one per cent. It shows that to the extent caste is co-terminus with class, there is inter-caste variation, at least on some of the items of expenditure. At the same time it also shows that, in general, irrespective of the caste affiliation, the migrants move out under distressed conditions and thus leading to homogenisation of the pattern of expenditure. This becomes further clear in our class analysis of the pattern of expenditure.

Landless agricultural labourers and marginal peasants spend major chunk of their earnings on the items of daily consumption. Expenses on health, repayment of debt and consumer durables are the other significant items of investment by them. The situation radically charges when we look at the preferences for spending savings by the well-to-do farmers, occurring more than 10 acres of land. More than half of their earnings goes into the medical treatment. Another 36 percent is spent on buying land or house or is invested in their development of agriculture.

Expense on daily consumption, which was highest for the labouring poor, was very low in case of bigger landholders. It can, therefore, be conducted that the rural poor masses leave their villages under distressed conditions and, often, do not possess money even to meet travel expenses. For this reason they borrow money on which they pay exorbitant interest rate. It is these poor masses who are forced to squeeze some money out of their wages which is largely spent on the items of daily consumption in the family. They also spend a substantial amount on medical treatment and out of the rest party goes in the clearance of debt and the remaining in buying consumer durables. The pattern of expenditure by the rural poor migrants suggests that they are caught into a vicious cycle of deprivation and pauperisation. Unless their savings become large enough to invest in the productively their poverty would breed only further poverty.

The foregoing discussion shows that during the last two decades the intensity and pattern of migration in rural Bihar has changed. The proportion of migrants has approximately doubled and has become widespread. Besides this, the nature of migration has changed from short term to long term. This is attributed to the fact that unlike in the past now workers constitute nearly two third of the total migrants. The migration of workers is fairly distributed across all castes, classes but as a whole upper castes and Muslims report more long-term migration compared to others. Many people, particularly from upper castes, now also prefer to work outside as it means that they can slowly break the existing caste taboos that exist in the village. While the upper caste people do not do any manual wage work in their villages because of caste taboos, they undertake also variety of work—wage work or loan-paid self-employed work—in their place of migration. This explains an important reason of upper caste youth migrating in large numbers. In case of lower castes also many migrate to come with from the clutches of the prevailing caste discrimination and exploitation at the hand of employers.

The caste and class analysis of migrant workers shows that while workers from lower caste and class are resorting to migration as an 'escape from exploitative employer' workers from upper caste and class are migrating for economic benefits. However, in both the cases there has been increase in migration of longer duration. In fact the nature of occupation at destinations, workers from lower caste and class are involved in, requires longer stay of migrant workers at the destination. Migration in these cases is not seasonal in nature.

In terms of choice of destination, migration is now more widespread. During the early 1980s the most important destination was rural areas of Punjab and Haryana. By the end of 1990s the highest concentration of migrants is in Delhi. Besides Delhi, migrants are now opting for many other town and cities as their destination. In fact, there has been a general change in the destination of migration from rural-rural to rural-urban.

The overall dependence of migrants on middlemen/agents has reduced over the years, and people have started to migrate of their own, most of the lower caste and class workers are still migrating through different private agents and middlemen. In case of migration through middlemen migrating workers are subject to extreme exploitation both at the hands of middlemen and employer.

Workers from lower caste and class are largely illiterate and or less educated and are absorbed low quality occupations in urban informal sector. Although the proportion of households reporting migration and receiving remittance is higher among higher castes and classes, remittances constitute as comparatively larger proportion of total household income of lower caste and class households. Nearly one-third of the total household income of the migrant household’s contributed by remittances—the proportion is much higher among landless and small landholders. This shows the dependence of households belonging to lower castes and classes on remittances.

Inter-State Migrant Workmen (Regulation of Employment and Conditions of Service) Act, 1979 is meant for regulation of the service condition of the migrant workers and improvement of their lot. There is need for making it more meaningful and stringent and to implement it sincerely and forcefully. It must cover both the origin and destination of migration.

At the origin of migration a record should be prepared about the prospective migrants and the process and destination their migration. No doubt there are practical difficulties in doing so. If they migrate through labour contractor such record can be prepared without much difficulty. In view of the increasing role of labour contractors and other middlemen, enforcement of some authorised agencies to mitigate or even eliminating their role may be useful in order to save the migrating workers from different kinds of exploitations. In other situations help of village level functionaries and originations such as VLWs, Panchayats or NGOs can be sought for this purpose. Such records will help in keeping a track of the well being of the migrant workers.

The provisions of the above mentioned legislation (Inter-State Migrant Workmen (Regulation of Employment and Conditions of Service) Act, 1979) regarding provision of free medical facility to the migrant workers, protective clothing, suitable accommodation, journey allowance etc. should be properly specified and defined and must be guaranteed to all the migrants. The other labour legislations regarding the maximum hours of work, minimum wage, child and women labour should strictly be enforced. The migrants should be also informed about the legislations meant for them and their rights, which they must get.

Proper facilities for the care and education of children of the migrants should be created at the destination of migration.

Since low economic development in general and inequality in the development of the villages and the adjoining areas in particular are the most important reasons of high incidence of migration from this region, the solution for it lies in rapid economic development of the state. Massive efforts are required for the development of the region in general and its villages in particular. Provision of cultivable land and assured irrigation in the villages, development of multiple cropping and improvement in the agricultural productivity may reduce the incidence of migration especially the seasonal migration from this region.

There is need to generate employment opportunities in the village. Development of small and cottage industries in the rural areas, promotion and protection of traditional artisan work and handicrafts will also create employment opportunities in the villages and thus reduce migration. Effective and extensive implementation of government measures for creating employment for rural poor like Jawahar Gram Samridhi Yojna (JGSY), Swarna Jayanti Gramin Swarojgar Yojna (SGSY) etc. will also reduce migration by creating additional employment opportunities for the prospective migrants. The NGOs, through promoting SHGs can also work in this direction.