1 Introduction

The desirability of internal migration and migration-development positivism reflected in the 1950s and 1960s emerged in the works of Lewis (1954) and Ranis and Fei (1961), who viewed rural to urban migration as a process of shifting surplus labourers from the rural ‘traditional’ sector to urban ‘industrial’ sector, where marginal productivity of labourers was not only positive but increasing (Todaro 1980). Return migrants were viewed as agents of change and innovation during that time (De Haas 2010), resulting in a win-win situation for both the source and destination areas. Neo-classical macro- and micro-theories of migration, within the broad functionalist paradigm, see migration as an optimisation strategy (De Haas 2021). Within this framework, migrants flow from low-market-wage areas to high-market-wage areas. In a parallel manner, since the capital-scare areas provide a comparatively higher rate of return, capital reverse flow to those areas is in the form of investment and high-skilled human capital (Massey et al. 1993). Neo-classical macro-theories explain migration as being the result of geographical differences in endowment of labour, which is reflected in the differential wages (Lewis 1954; Todaro 1969). Rooted on experiences of the advanced industrial economies with near-full employment in urban areas, the experiences of third world countries were marked by rather slow urban job creation that could not accommodate the growing migration to the cities (Todaro 1969, 1980). Despite rising levels of urban unemployment and underemployment, city-ward migration continued in 1970s and 1980s. Internal migration at that time was viewed less sanguinely, both in terms of effects on rural productivity and income distribution (Lipton 1976 cited in Todaro 1980). Todaro (1969) postulated that migration decisions are based on expected earnings defined by employability and earning differences, rather than actual wage differences between the source and destination areas. A rational migrant moves when his prospective net return at the place of destination is positive. Against the assumptions and conclusions of neo-classical theories, the new economics of labour migration (NELM) theories of Stark and Bloom (1985) consider household units and families as the decision-making unit, whose incentives to migrate come from relative deprivation rather than absolute income. Under NELM, households try to diversify their income by allocation of family labour to reduce any foreseeable income risk (De Haas 2021). Remittances received by households are important to avert risks in developing countries where institutional mechanisms like private markets and government policies are relatively not well developed (Stark and Levhari 1982; Katz and Stark 1986; Lauby and Stark 1988). Despite all shortcomings, functionalist theories have great explanatory values to explain labour movement within an internal and restriction-free environment.

The concept of return migration is not new in the migration framework. Ravenstein (1885) postulated a counter-current to the mainstream of migration in his pioneering work on formulating migration laws. Subsequently, Lee (1966) postulates that economic conditions affect the efficiency of migration, as migration is not one way, but during times of depression, migrants may return to their sources. The neo-classical and NELM theories conceptualise return migration very differently. In the context of wage differentials that motivate migrants, the neo-classical theories have difficulty explaining return migration because of their underlying assumption that a migrant’s decision to migrate is a well-informed one, and so there is no rationale for a migrant to return. The later variants and extensions of the theory, however, took account of the expected earnings, which are calculated as the product of real wage and employability at the destination (Todaro 1980). Return migration, on the other hand, is viewed by the NELM as a calculated strategy for migrants after they have achieved a certain goal at their destination and as a natural outcome of the migration process (Stark and Levhari 1982; Katz and Stark 1986; Lauby and Stark 1988). Both these sets of theories have fundamentally different ideas about integration with destinations and ties with origin areas. Within neo-classical framework, migration is a static process, with the purpose of migrant individual income maximization. Ties with source areas, be it emotional or monetary, increase the cost of migration (Massey et al. 1993). In this context, NELM theories differ in that families and migrants use migration as an income diversification strategy, with remittances being used to both invest and protect the household from monetary risk.

There has been burgeoning research into the traditional success–failure dichotomy of return migration. Previous literature on determinants of return migration has found evidences of greater integration with destination, as evidenced by marrying the locals, speaking their language, and longer stays reducing intension of return migration (Dustmann 1996), socio-cultural integration with the destination has a negative effect and transnational ties have positive effects on return migration from four African immigrant groups in Spain and Italy (De Haas and Fokkema 2011), social considerations among Swedish immigrants (Niedomysl and Amcoff 2011), higher investment and social ties with sources, and lesser socio-cultural integration across Europe among immigrants of Morocco (De Haas et al. 2015) were all found to be significant. Another approach of testing the success-failure theory is to see if the return migrants are more able or more advantaged than non-migrants (Wang and Fan 2006). On the characters of returnees, it is accepted that return migration is a selective process, where returnees can be either positively or negatively selected and selectivity varies by groups of migrants (Constant and Massey 2002). In their study among internal rural–urban migrants, Wang and Fan (2006) found that with increasing age the likelihood of return increases. Apart from age, men, married individuals, and size of arable land at homes are positively related to return propensity among migrants, while training in non-agricultural work and education are negatively related (Wang and Fan 2006). Poor health is significantly related to return migration among Mexican immigrants to the USA (Arenas et al. 2015). A recent study based on secondary data reports two categories of return migrants to India, the poor and marginalized and the highly educated ones (Dhar and Bhagat 2020). In the migration-development nexus, the nature of the return migration is another dimension to consider. Return migration can be voluntary or involuntary and forced, as well as temporary and permanent in nature. Both the neo-classical and NELM theories don’t provide this distinction. Apparently, the basic assumption is that return is voluntary in nature (Mensah 2016). Literature pertaining to return migration is skewed towards voluntary streams and how they contribute to the economies of receiving areas. The majority of studies on involuntary or forced return migration focus on asylum seekers and refugees (Ruben et al. 2009; Drotbohm 2011). Involuntary return can be identified with characteristics like—‘forcefully expelled’ migrants to their home countries (Ruben et al. 2009), don’t have a preparation period before leaving (Mensah 2016), and the absence of two the elements: freedom of choice and an informed decision of the migrants (IOM 2012 cited in Mensah 2016). Return migrants, particularly the involuntary ones, are more likely to be in a stage in the migration process where they don’t necessarily want to resettle and are thus more mobile (Anarfi and Jagare 2008). Based on evidences from the previous literature, it is understood that there is large ambiguity in the motivation, process, and consequences of return migration and characteristics of return migrants.

2 The Pandemic, Resultant Lockdown, and Return Migration

The outbreak of Severe Acute Respiratory Syndrome (SARS) Coronavirus-19 started as a health crisis and later, rapidly turned into a devastating humanitarian and social catastrophe. India’s first case was confirmed on 30 January 2020 in Kerala (Perappadan 2020). This state from South India saw an early spike in cases. However, as a result of numerous efforts, the numbers quickly came down. By the end of April 2020, Maharashtra, Gujarat, and Delhi—the states with the highest per capita income, more urbanised, industrialised, and destinations for a large share of country’s labour migrants—were the worst affected ones (Ray, 26 May 2020). As an effective safeguard in responding to the virus, India declared an early lockdown on 25 March 2020 for 21 days in the first phase. Economic indicators began to show declining trends soon after. According to a study conducted by Kesar et al. (2021) among 5000 respondents across 12 states in India in April and May 2020, two-thirds of workers were losing their jobs. The same study found a significant drop in wages among workers who continued to be employed during the lockdown. According to CMIE data, the prevailing rate of unemployment at the time of the lockdown increased threefold and reached up to 27% (The Hindu, 05 May 2020).

As a result of the nationwide lockdown, the vulnerabilities of the migrant workers working in the urban informal sectors without access to social security were exposed in the most horrific ways. To make the matter worse, Indian railways had decided to suspend all passenger trains starting from 23 March 2020. The railways had started the specials on 1 May 2020 for migrant workers hit by the pandemic and resultant lockdown. In the meantime, some migrants footslogged hundreds of kilometres to reach their native villages. According to Indian Railway estimates, approximately six million labourers use the special service of ‘Shamrik Trains’ during the lockdown (The Hindu, 15 June 2020). The movement of migrant labourers, both internal and external, back to their source areas due to the crisis is identified as a ‘reverse migration’ (Dandekar and Ghai 2020; Premkumar 2020).

There is no systematic study on return migration in India (Dhar and Bhagat 2020) indicating a major gap in knowledge on migration. India’s internal labour migration during the pandemic provides a unique experience of involuntary return migration. Although, of late, the situation has been changing, earlier migration used to be viewed as a failure in development and often was discouraged in India (Bhagat 2017). The paradox of migration-pessimism on the face of growing concern about India’s low rate of urbanisation has received academic attention (De Haan 1997). The reason for pessimism for internal rural–urban migration may be because of growing problems in cities such as urban poverty and slums. Previous works have examined the ‘spill-over’ effect of rural poverty to urban areas, most notable of which are Dandekar and Rath (1971) and Mitra (1992). Additionally, there exists a ‘receiving country bias’ in migration literature that hinders the generalised understanding of migration (De Haas 2021). The pandemic and selective return migration of labourers has put the focus back on the complex migration process as an integral part of development, forcing policymakers to reconsider the issue from the perspective of migrant labourers engaged in the most precarious jobs. At this juncture, the pertinent question is whether migrants place a burden on the source areas or present a window of opportunity for the local economy to capitalise on. In this context, it’s crucial to know who returned, what their backgrounds are, their ties to the receiving and destination areas and their residential quality of life, and what their current employment status and future aspirations are.

3 Theoretical Framework and Data

3.1 Data

This study is based on telephonic questionnaire-based survey done in Sonitpur District of Assam. A total of 238 people were surveyed between February and March of 2021. The contact numbers of migrant workers were collected from Presidents of several Panchayat Samiti in the district. A list of out-migrants from respective development blocks, who were still away, and those who had returned home during pandemic was availed. One of the main drawbacks of the sample is that we were unable to compile a list of migrant labourers from all development blocks. We did not apply quota based on the destination or source villages of migrants. We tried to contact around 500 individuals selected at random from the list of migrants, of whom 238 responded to us and participated in the study. The participants of the survey came from various social groups, including the adivasi (tea-tribe), who make up roughly 12% of the total sample. The sample may be biased towards the non-returning migrants. Furthermore, due to a high non-response rate, the respondents may not be representative of the migrant population from the area. After the lockdown was declared, 156 out of the 238 (65.8%) returned to Sonitpur District. At the time of survey, 109 of them were still living in their homes, while 47 had returned to their previous destinations. The remaining 81 (34.2%) individuals stayed back in their destination areas during lockdown. All 156 individuals who returned to their native places after the declaration of nationwide lockdown are considered return migrants.

3.2 Study Area

Situated on the north bank of river Brahmaputra, total population of Sonitpur district is about two million, according to the 2011 Census. In terms of inter-district migration within Assam, Sonitpur is a positive net-migrating district. The population density is 370 persons per square kilometre. Decadal growth rate of population in the last decade is 15.6%. Overall literacy rate is 68.6%. 10.8% of total population lives in urban areas.

3.3 Determinants of Return Migration

Amidst the pandemic and nationwide lockdown, not all migrants returned home, some stayed back. We determined returning based on their response to the question, ‘did you or did you not return back to your native places with the intention of staying there after the lockdown was imposed?’ If they answered yes, they were classified as ‘return migrants’, irrespective of their length of stay. Since our dependent variable is dichotomous, we have performed binary logistic regression to find out determinants of return migration. The sample is further reduced to 229 cases due to missing values on one or more of the relevant variables.

We draw the research framework on existing literature on return migration during the time of crisis, notably in the context of economic crises. A substantial body of literature is available on cross-country return migration due to economic reasons, primarily due to job loss during the Great Recession. A substantial portion of it focuses on the experiences of Europe, and North America, while some are also from South Asia (Zachariah and Rajan 2010; de Hass et al. 2015; Zaiceva and Zimmermann 2016). Evidence shows that, during the trouble time males, single, older migrants, migrants without children, and those with higher or middle educational attainments are also more likely to return (Zaiceva and Zimmermann 2016).

‘Reverse migration’ during a pandemic and return during economic crises are similar in many ways, but they are not the same. First, evidence suggests that cross-border migrants are less responsive to economic crisis and choose to remain in their host countries (Pusti 2013). This is because, during economic recessions and depressions, migrants' economy in their home countries suffers just as much, if not more, than those in their destination countries. Pandemic and lockdown elicited strong reactions among migrants. Second, Covid-19 began as a health crisis and rapidly evolved into a humanitarian catastrophe. As the crisis evolved, economic activities became limited, and mobility has been restricted, providing a barrier to accessing necessary services such as health care and daily supply of ration. Migrant labourers stay away from the comfort of the family environment on a meagre income and limited savings, and their lives were put in jeopardy owing to the labour market shock. On top of all, it was becoming increasingly difficult to predict when the situation would improve. While return migrants from big cities were viewed less sanguinely in the countryside, many migrants resorted to desperate tactics in dire times; thousands of migrants walked the full distance to source locations in the countryside. Despite all the mistrust, the severe financial crisis and such a catastrophic scenario prompted an urgent yearning to return home, particularly for those without job stability and insufficient means of staying at the destination. As Covid-19 cases sore in the cities in the beginning, countryside remains nearly untouched, providing still another reason to return. People of all classes were affected by the pandemic and the ensuing lockdown; however, poor migratory labourers living on minimal wages and lacking basic amenities in their destination are likely to suffer the most. Availability of housing and food and the availability of private transportation are used as proxy for residential quality of a migrant's destination location. The quality of residence shows how well a migrant is doing in his or her area of destination. A person who is well-equipped is less likely to return.

Migrants’ return decision can also be explained by the presence of economic alternatives such availability of land, having a job card, and partner’s engagement in economic activities. The logistic regression model includes these factors. Factors related to ties with origin areas act as pull-factors in the return process. Migrants’ ties with sources are sustained through social contacts, by occasional visits and through remittances sent. Migrants having stronger ties to their sources are more likely to return. The percentage of monthly income returned home as remittances is used as a proxy for ties with source areas.

Return migration decisions, according to neo-classical theories, are associated with failure of integration with the destination locations. Integration with the destination is a framework derived from international migration. Moroccan migrants' decision to voluntary return from Europe and their socio-cultural integration with host countries has a negative association (de Hass et al. 2015). In another study, evidence from Spain further suggests that the migrants bring families to the host countries or families are formed in the host countries which strengthens social ties with the destination. This gives the migrants a reason to prolong their stay in the destination during a crisis (Pusti 2013). Integration with the destination demonstrates a sense of belonging and security (Pusti 2013). It is reasonable to believe that migrants use such ties for economic and emotional respite in the face of pandemic-induced insecurity. The logistic regression model includes three indicators linked to integration with host areas: families migrated with the migrants to destination locations, the nature of migration (annual or seasonal), and the number of languages spoken by migrants. Seasonal migrants, migrants who speak fewer languages and migrants who left their families behind are more likely to return.

The logistic regression model includes other socio-economic characteristics such as income, migrants' age and gender, educational attainment, social categories and marital status in addition to these sets of variables associated with integration with destination, ties with sources, and residential quality at destination. Participants of the study are categorized in to three social categories: the Adivasi, Nepali, and others. Adivasis are part of the tea-tribe population, which is classified as More Other Backward Class (MOBC). Nepali out-migration from Sonitpur district has received academic attention (Muktiar and Sharma 2019). The Bodo and Missing tribes, Bengalis, and everyone else not listed in the first two groups are considered 'others.' The possibility of multicollinearity among the explanatory variables is ruled out as the Variance Inflation Factor (VIF) score is below three.

4 Migration in North East India

Census 2011 recorded little over a million out-migrants from the seven sister states, viz. Arunachal Pradesh, Assam, Manipur, Meghalaya, Mizoram, Tripura, and Nagaland, all located in India’s North East. They constitute 2.3% of the population of the region. However, about half of these out-migrants moved within the Northeast region. Assam is an exception; it is a net out-migrating state, with more than 50% of all out-migrants heading for rest of India. Tripura and Manipur, together with Assam, are the other two net out-migrating states in the region. Assam and Tripura receive a considerable volume of immigration, and thus loss of population due to internal out-migration is offset by population gains by immigration. Based on Census data on migration by place of birth for 2011, four remaining states of the region, namely Arunachal Pradesh, Nagaland, Mizoram, and Meghalaya gained population due to internal migration. Among those who migrate outside the region, the urban-to-urban stream of migration predominates, with 52% of out-migrants coming from urban areas and 68% of all out-migrants from the northeast going to urban areas. About 23.3% of all out-migrants move out to six urban agglomerates: the National Capital Territory, Kolkata, Greater Mumbai, Hyderabad, Bangalore, and Chennai (Lusome and Bhagat 2020). Numerous recent studies have pointed out a new trend of migration to South India, most notable among students (Douminthang and Xavier 2020), construction workers (Narayana et al. 2013) and female employees in beauty-care services (Deori 2016; Deori and Rajagopalan 2017). Southern states are booming as the new destination for out-migrants from north east India because of several reasons, including higher wages, a more welcoming society and faster industrialization, to name a few (Lusome and Bhagat 2020).

Census data from 1981 show a steady increase in out-migration, notably among the youths of the regions, which is noted in scholarly work (Chyrmang 2012). A variety of 'push factors' are cited to explain this trend among youngsters leaving the northeast, particularly, for ‘employment and education considerations’ (Remesh 2012). The double whammy of flooding and erosion is gradually lowering the man-land ratio to man's detriment, and the area's long history of insurgency, low agricultural productivity and sluggish industrial growth all contributed to out-migration (Shimray 2004; Remesh 2012; Muktiar and Sharma 2019). According to a recent study, economic reasons are more efficient in influencing out-migration from the region than social reasons (Bhowmik 2021). In contrast to previous studies focusing more on Delhi and other metropolitan cities, there has been a noticeable shift in recent studies in terms of migrant destinations. The current streams of migrants have shifted their focus to the southern states, as reflected in the recent findings (Narayana et al. 2013; Douminthang and Xavier 2020; Deori 2016; Deori and Rajagopalan 2017).

The present-day states of Assam and Tripura have historically been receiving states in the region. Internal migration began during British rule as a result of the steep population density gradient towards the erstwhile East Bengal, and resource endowments such as lush land and tea estates in Assam (Bandyopadhaya and Chakraborty 1999). A negligible portion of in-migrants went to the remaining peripheral areas, such as the erstwhile hilly districts of Assam. Additionally, the British imposed the Restrictive Bengal Regulation Act of 1873, or ‘inner line restrictions’ to protect tribal identity and culture, which is another reason why major chunk of in-migrants went to districts located in Brahmaputra plain and Barak-Surma valley of Assam (Dikshit and Dikshit 2014).

5 Results

5.1 Destination Regions and Background of Migrants

Migrants from Sonitpur have a strong predilection towards Southern India; four southern states namely Karnataka, Maharashtra, Tamil Nadu, and Kerala, together constitute 78.6% of all out-migrants from Sonitpur District of Assam. Among the four states, Karnataka alone attracts 54% of total out-migrants. The result corroborates with broader pattern of out-migration from Northeast India to South India, as shown by some of the earlier findings based on Census figures (Reimeingam 2016; Lusome and Bhagat 2020). Apart from those four states, people also migrate to Delhi and West Bengal. About 8% of out-migrants move to Delhi, whereas 3% move to West Bengal. Among Indian cities, Bangalore is the favourite destination, accounting for half (50.4%) of the total out-migrants, followed by Delhi (7.6%), Mumbai (5.9%), Pune (2.9%), and Chennai (2.1%) (see Figs. 1, 2).

Fig. 1
figure 1

Source: Primary Survey, February–March, 2021

Destination states/country among out-migrants of Sonitpur District, Assam.

Full size image
Fig. 2
figure 2

Source: Primary Survey, February–March, 2021

Destination cities among out-migrants of Sonitpur District, Assam.

Full size image

The majority of the migrants are young (86% are between the ages of 20 and 40), the median age is 28 years, single men (53% unmarried), and without a high school diploma (75%) who migrated looking for work (see Table 1). The median income of the workers is 12,000 rupees a month (155 US $). Only 5.6% of the people in the sample migrated for non-economic reasons, such as moving with family or to pursue higher education. The remaining 94.6% were mostly engaged in the urban informal sector, with some level of regularity in salary albeit without social security. Most migrants move in a group, aided by a network of relatives and acquaintances with previous migration experiences. They assist the new ones in finding jobs in the destination places, forging their own communities in the process. At the time of crisis, most labourers returned back in groups. Often, we observed, occupational segregation among migrant labourers by community. For instance, Nepali migrants with surnames like Phuyel, Newar, Bhujel, and Dhakal were often seen working in hotels and restaurants. However, there is little evidence to conclude that community-based labour segmentation is foolproof; informal networks among migrants may foster such selectivity and may not reflect any class or caste division as such at work.

Table 1 Description of dependent and independent variables
Full size table

5.2 Economic Activities of the Migrants

The majority of the migrants worked in precarious jobs in the unorganised sector. There were specific occupations (based on National Classification of Occupation 2015) where migrant workers were largely absorbed, including the hotel line as waiters (34.2%), gateman/ unarmed security guards (12.8%), building construction labourers (6.8%), building housekeeping (4.7%), street food vendors (4.7%), commercial vehicle drivers (4.7%), and stall and market salesperson (3.8%). Majority of these occupations are annual in nature, providing migrants with a regular flow of income throughout the year, while just a very small percentage (4.2%) are seasonal in nature. A quarter of migrant workers (25.3%) have a seven-day work routine; 73.9% work 6 days a week, and the remaining workers work 5 days a week or have no fixed work schedule. The bulk of the hospitality industry workforce (waiters, cooks, and housekeeping) were provided food and accommodation at their workplaces. We discovered from our interactions with workers that the benefit of rent-free living comes with a catch: the extra workload. Those in the hospitality business labour late at night during busy days. Some workers in the courier and delivery services receive incentives in addition to monthly income for working extended hours, late at night, and on holidays.

Since lockdown was imposed, the labour market has deteriorated. The massive unemployment and decline in average income figures claim that migrant labourers were hit the hardest. Several establishments had closed, some employers refused to pay, and some workers were forced to quit as the number of customers declined. From February to March, businesses ceased paying salaries. 13 (5.5%) migrant employees stated that they were not paid for the last month they worked before returning home. Return migration began shortly after the lockdown was imposed on March 2020, with roughly 2.5% of migrants returning. In the following two months, May and June, a massive exodus from various parts was seen, when 31.8 and 18.5% of all return migrants came back to their native villages. Despite the fact that wages had ceased by March, a migrant worker who previously worked in hotels in Trichy reported receiving free meals from their employers as an incentive to stay and that the fear for their lives had compelled them to return. We estimate that, over the subsequent 12 months till the time of the survey, about 30.1% of return migrants out-migrated. One-fourth (28.4%) of the remaining returnees don’t want to go back to their previous location either due to lack of motivation or insufficient motivation and are seriously considering a new life at home. They have myriad reasons, the most frequently mentioned of which are dissatisfaction with their work and city life, a single male household, and anxiety and fear about another similar situation in future. Migrants learn new skills while they are away; one Kerala returnee said he learned new skills after working in the construction industry for more than a year, and he plans to employ them in his native place. Skills learned while being away also aid migrants in securing a livelihood once they return home. The other section of migrants, based on their migration aspirations ranging from involuntary migration to voluntary economic migration, is a mixed lot: some have not made up their minds regarding returning back, some are reluctant to move at the same time don’t have any option to fall back on, some are willing to return and planning for the perfect opportunity, and some are dissatisfied with the conditions at home and want to re-migrate very soon.

In this section, we compare income and employment between the two time periods, i.e. before and after the lockdown. We looked at the unemployment figures from February and March 2021 to a get sense of the devastation about a year after the first lockdown was imposed. Over a quarter of migrants (27.5%) lost their jobs during the last 1-year period, and a large portion of those who were employed had their incomes decline significantly. Workers across all age groups have experienced job loss. 29.5% are still jobless a year after the lockdown at the time of the survey. Furthermore, job losses were observed across all income levels except for the extremely poor categories, whose monthly income is less than five thousand a month, increased to 7.6% (see Fig. 3).

Fig. 3
figure 3

Source: Primary Survey, February–March, 2021

Income of the migrants before and after the pandemic, Sonitpur, Assam.

Full size image

The findings of the study suggest that the majority of migrants are single men who leave their families at home to seek work in the big cities. Only 16% of the migrants migrated with their families. During their stay away, they kept a close tie with their families by visiting them and sending remittances to supplement income at home. In the entire sample, nine out of 10 out-migrants sent remittances to their families. This provides strong evidence in support of NELM to claim that migration is an income-diversifying strategy for poor households, and migrants remit to sources, thus maintaining close ties with the sources. Migration is a survival strategy for poor households without physical assets such as cultivable land. Only 10% of migrating households have cultivable land, and even among those who have, the land holdings are small. After their return, in the absence of income from jobs, a small share of them are working as a daily wage labourers in construction sites and tea gardens, that very livelihood that once had forced them to migrate to cities, while semi-skilled workers continue to work as tailors, motor mechanics, and car washers. Some others have taken up broiler farming and vegetable farming in their homes with the remainder of their savings. More than three-quarters of migrant households (77.1%) have job cards.

Internal out-migration is particularly acute among young single males without much education, most of whom are moving without families to big cities with a strong desire to acquire a livelihood which is the general characteristics of out-migration from Sonitpur. Only a small percentage of migrants migrated for the purpose of education and business. These characteristics point to economic deprivation that spurs distress migration.

5.3 Determinants of Return Migration

Return propensities are higher among males than females, among married migrants, among migrants whose family owns the land, among migrants who did not receive last month's salary, among migrants with a high school diploma, among seasonal migrants, among migrants whose family did not migrate, and among migrants who were not offered accommodation by their employer, according to descriptive statistics (Table 2).

Table 2 Determinants of return: descriptive of return migrants
Full size table

The logistic regression results in the form of odd ratio and significance level are provided in Table 3. Social and demographic variables such as age, sex, education, and partner's job status do not appear to have a significant impact on the decision to return. Return migration decisions are not influenced by factors related to quality of life at destinations. After controlling for all variables, it was revealed that the percentage of income sent as remittances, the availability of a job card by migrants' households, status of family migration, migrants’ income, and the number of working days per week are all significantly related to migrants' decision to return. Return decision is positively associated with the percentage of income remitted by migrants. Migrants from households with a job card are 3.26 times more likely to return home. Migrants who work 6 days a week are twice more likely than those who work 7 days a week to return. A significant portion of migrants who worked 7 days a week worked as waiters and cooks in hotels and restaurants. Following the lockdown, participants reported being offered with free food and accommodation as an inducement to stay by their employers, which may explain their low return. The income of the migrants has a significant positive association with their decision to return, but other variables are more important.

Table 3 Determinants of return migration
Full size table

6 Conclusions

The majority of research on the factors that influence return migration is centred on migrants' proclivity to return to their home countries after living in another country. The SARS coronavirus-19 created a unique situation of involuntary return migration of workers within the country. At the same time, it provided a stage for an opportunistic retrospective study of the factors that influence internal migrants' return behaviour. The purpose of this paper is to conduct an empirical verification of migrant workers' socio-economic profiles and destination information, analyse factors of return migration, and comprehend return migrants' future aspirations.

We convey responses based on a quick survey and data collected from Sonitpur District of Assam. Four states from South India are favourite destinations among out-migrants. Majority of the migrants are young, single males with little or no education, and they leave their families behind to look for work in big cities. In the absence of assets like cultivable land, most migrant households resort to remittances sent by the migrants. About nine out of ten migrants send back remittances to home. Among those, who had returned, one-third migrants already went back and a significant portion are willing to go back to their previous destinations. Keeping this in mind, the intervention should not only be aimed to stimulate further return; it should focus on the success of re-integration process by providing financial help and skill development that is required to make a new start and necessary living.

Several variables are included in this study, including job card and land availability, language spoken, and nature of migration, as well as indicators important in the context of pandemic-related return, such as quality of life in destination areas as revealed by the status of accommodation and food, and transportation, and the number of work days per week, which is otherwise overlooked. The regression analysis findings indicate substantial evidence in support of the new economics of labour migration, which states that migrants maintain strong relationships with their origins and that those with stronger ties are more inclined to return during times of crisis. The findings also corroborate the neo-classical theory hypothesis that the inability to integrate is linked to the decision to return. The presence of a family in the host place aids in the formation of strong ties to the destination. The findings of the study reveal that single migrants are more likely to return. The hypotheses of both sets of theories, the neo-classical and NELM are complementary in explaining the return process. Finally, we suggest a hypothesis based on the observations that during times of crisis, migrants with other economic options at sources, such as a job card, are more likely to return.