Understanding the skills needs of migrant workers from Nepal Uttam Sharma1 Maya Sherpa2 Sangeeta Goyal3 Abstract Migration abroad is an important source of work and income for Nepali workers. In a study of current and returnee Nepali migrants (outside India), we find that migrants, who are largely male, in their mid to late twenties and with relatively low levels of education, express the need for and are willing to bear the cost of occupation-specific skills training. The average earnings of migrants in the sample with training were considerably higher compared to those without training. Migrants were mainly engaged in low- and semi- skilled occupations in the destination countries (outside India), such as laborer, construction worker, security guard, driver, and the skills needed for these occupations were acquired informally. Few migrants had undergone any formal skills training prior to their migration episodes. The areas in which training was desired were semi-skilled occupations (driver, electrician, cook, computer operator, and so on), and the modal duration of desired duration of training was three months, followed by six months. The findings from this study suggest that there is an unmet need for occupation-specific skills training for migrants. This need has become more urgent for those who have returned to their home country with the restrictions on employment abroad imposed by the pandemic. 1 World Bank, uttamsharma@gmail.com 2 World Bank, msherpa@worldbank.org 3 World Bank, sgoyal2@worldbank.org 1 1. Background In search of better opportunities for employment, a large number of working-age Nepali citizens have gone abroad for employment. Remittance income from migrants constitute more than 25 percent of the national gross domestic product (GDP) since 2012, reaching as high as 31 percent in 2015 and 2016.4 In fiscal years 2017/18 and 2018/19, a total of 590,306 people obtained labor permits from the Department of Foreign Employment (DoFE) to be eligible to work abroad, not including India.5,6 The same report also says that, nearly 60 percent of labor migrants in 2018/19 were deemed unskilled and face higher economic costs and risks of migration compared to semi- and high-skilled labor migrants. The share of unskilled migrants, however, decreased by 5 percentage points between 2017/18 and 2018/19. Additionally, there is national concern over the integration of returnee migrants in the domestic economy in productive ways, which now has become an immediate priority for the Government of Nepal (GoN) with a large number of these migrants returning due to the COVID-19 pandemic. This paper explores answers to various migrant-related questions in the context of Nepal. In particular, it tries to understand the occupation-specific skill7 needs of migrant workers, both current migrants (who are presently abroad for work) and returnee migrants (who have worked outside Nepal but have not made any plans to go for work in the near future). It also considers the perceptions of both categories of migrants including how they perceive different policies that are or can be implemented to help migrant workers. For example, we explore their views on testing and certification of skills acquired abroad as it has the potential to enable returnee migrants to signal their higher skills to employers and facilitate further formal upgrading of their skills in the future.8 The strength of this study is that it asks detailed questions on the skill needs of migrant workers, both current and returnee migrants. In addition to in-person interviews with returnee migrants, family members of current migrants and current migrants temporarily in Nepal, phone interviews with the current migrants residing abroad were also conducted. All seven provinces of Nepal are covered in this study. There are two potential limitations of the study. We have tried to identify the skill needs from the perspectives of the migrants. This study, however, does not include perceptions of firms abroad employing Nepali workers. One could argue that skill needs in recipient countries keep changing and the firms abroad would be better placed to identify the emerging skill needs. The second limitation of this study is that it 4 Baniya, Jeevan, Sadikshya Bhattarai, Binay Jung Thapa, and Vibhav Pradhan. 2020. “COVID-19 and Nepali Labour Migrants: Impacts and Response. Center for the Study of Labor and Mobility. Remittance income includes earnings sent home by all migrants, including those who find work in India as well as those who migrate without the need to be issued labor permits. 5 Labor permits are not required for Nepali citizens to work in India. 6 Ministry of Labour, Employment and Social Security. 2020. “Nepal Labour Migration Report 2020.� Government of Nepal. 7 ‘Occupation-specific skills’ and ‘skills’ will be used interchangeably in the paper. The paper follows the definition of skills by the National Skills Testing Board (NSTB) which is the central agency certifying short-term skills in Nepal. 8 The survey was carried out in close cooperation with the Project Secretariat, the implementing agency of the Enhanced Vocational Education and Training Project II (EVENT II), which is a GoN-implemented project whose objective is to expand the supply of employable labor through improved access to market relevant training programs and strengthened Technical and Vocational Education and Training (TVET) sector service delivery in Nepal and other institutions working to support migrants. 2 does not adequately discuss the ‘how’ part of the problem. Though it details what skills are needed by the potential migrants, it largely remains silent on how to practically train the migrants in the skill needs in recipient countries. The major impediments to the conduct of training are that migrants are located in diverse areas across Nepal and are present in the country at different times. To understand the skill needs of migrant workers, it is important to capture the basic profile of the migrants, including what they are currently working on. After going over the sampling strategy and data collection in the next section, Section 3 first provides the profile of migrants, including countries they have worked, age, gender, marital status, and their education level. We then present findings on migration experiences (countries and number of episodes), sources of information on job placement abroad, job performed abroad and total earnings, and some regression results on correlations between migrant income and migrant characteristics, including the value added from having more education and training. The training and skill needs of migrants are considered in detail next. Further discussion and concluding thoughts are presented in Section 4. Section 5 discusses possible uses of data collected for this survey for integration of migrants returning to Nepal due to pandemic-related job losses in destination countries. 2. Sampling Strategy and Data Collection Since the objective of this work was to understand the skill needs of current and returnee migrant workers, it was decided to focus on 14 districts (two per province) across Nepal. Fourteen districts that had a large number of their population in foreign employment were purposefully selected, two from each of the seven provinces. The names of the districts and the provinces they represent are provided in Table 1. Table 1: Districts in the study sample9 Province District 1 District 2 1 Jhapa Morang 2 Dhanusa Mahottari 3 (Bagmati) Makwanpur Chitwan 4 (Gandaki) Nawalparasi Tanahu 5 (Lumbini) Rupandehi Dang 6 (Karnali) Surkhet Salyan 7 (Sudurpaschim) Kailali Kanchanpur Once the districts were chosen, the second step was to select wards10 from the abovementioned districts. To identify pockets of areas from where foreign employment migrants is high, wards with the highest number of households with foreign migrants, using the Nepal Population Census 2011 data, were selected for each district. This was done to increase the chances of finding households with either a current or a returnee migrant in that area. Each ward was then considered as a cluster or the primary sampling unit from which the households were selected. Across the 14 districts, 80 such clusters were generated with 25 households from each cluster, so that a total of 2,000 households would be interviewed. These clusters were largely divided across the 9 Provinces 1 and 2 have not been given names yet. 10 Wards are the smallest administrative units in the rural or urban municipalities 3 districts based on the proportion of their contribution to the foreign employment migrant population. In other words, more clusters were assigned to districts that had larger foreign labor migrant population. For the selection of households from the selected wards, a listing process to identify 40–50 households having a returnee—either a foreign employment migrant outside India in the last five years or a current foreign employment migrant outside India age 18 and above—was conducted. Within each selected ward, random start points were identified to start the task of listing the households meeting the requirement. From this list of households, 25 households from each ward were then randomly selected. The next step was to choose a respondent to the household survey component of the study. In case of households having a returnee migrant, the interview was taken with that individual regardless of whether that household had a current migrant. In case of the households without returnee migrants but having a current foreign-employed migrant, who was not temporarily in Nepal, the interview was taken with the household head or the person most knowledgeable about the current migrant. If there were more than one returnee migrants in the household, the one who was present in the household was selected. In case more than one current migrant was in the household, the one with the higher number of migration episodes in the last five years was selected. If there was a tie even with these conditions, the respondent was chosen randomly. In addition to the household survey, phone interviews with a subset of absentee household members who were abroad outside India were conducted. The phone interviews include calls made on Viber, Whatsapp, and Facebook messengers. Their contact details were obtained from the household survey respondents. Though the target was to interview randomly selected 500 absentee household members currently working outside Nepal and India, we ended up interviewing all 331 individuals who we were able to trace and were willing to talk to the field researchers from Solutions Consultant, a survey firm that was hired to collect data for this study in April–June 2018. The number of questions for the phone interview respondents was fewer than for household survey respondents, given the difficulties of conducting long interviews over phone. Current migrants who responded to our survey and who did not respond are very similar in characteristics such as age, gender, and number of migration episodes. Migrants from Malaysia are represented slightly more in the phone survey, presumably because it was easier to interview them in the evenings when most are relatively free. On average, slightly more educated people are non-respondents (8.5 years of education for non-respondents to 8 years of education for respondents). In total, 2,000 households with current or returnee migrants (in the last five years) from outside Nepal and outside India were interviewed. The details on number of households interviewed, number of households with returnee or current migrants, and whether phone interviews were conducted with an absentee household member residing outside Nepal and India is provided in Table 2. Note that of the 2,000 households included in this study, 167 had both returnee and current migrants. As a result, the sum of households with returnee migrants and current migrants abroad exceeds 2,000. 4 Table 2: Details on households and phone interviews conducted, by district Households Households Households with Name of Households in the S.N. Province with returnee current Districts interviewed phone migrants migrants abroad interview 1 1 Jhapa 250 106 157 73 2 1 Morang 250 189 74 27 3 2 Dhanusha 300 184 156 51 4 2 Mahottari 250 157 139 51 5 3 Chitwan 100 68 45 10 6 3 Makwanpur 100 59 45 12 7 4 Nawalparasi 175 102 90 39 8 4 Tanahun 100 64 42 4 9 5 Dang 125 107 53 19 10 5 Rupandehi 150 79 46 15 11 5 Salyan 75 35 15 9 12 6 Surkhet 50 59 19 5 13 7 Kailali 25 9 17 6 14 7 Kanchanpur 50 16 35 10 Total 2,000 1,234 933 331 A decision to focus this study on current or returnee migrants who have worked outside Nepal and India in the last five years merits further clarification. The profiles and views of both current migrants and returnee migrants are adequately reflected here, which allows us to have a comprehensive view of migrants’ needs. Since one of the objectives of this study was identifying skill needs of migrant workers, it was felt that recent migrants would be able to identify current trends better and be able to suggest next steps to help migrants in general. This means some relatively older adults who had returned to Nepal more than five years ago are not included in our sample. This study also does not include migrant workers to India, partly because the proximity and open border in place between Nepal and India means that such migrants may be very different from migrants who go for work to other countries. 3. Findings Profiles of migrants Before going over other descriptive statistics from both the household survey and phone interviews, we consider the profile of all the migrants outside Nepal and India in the 2,000 households that are a part of this study. We have listed the basic profiles of all the migrants that are part of these households. Their profiles are compared with phone interview respondents where possible. Of the 2,402 migrants from 2,000 households, 54.9 percent were permanently living in Nepal (returnee migrants); 4.7 percent of the migrants were temporarily in Nepal and would return to another country outside India for work; and the remaining 40.4 percent were currently outside Nepal and India for work. The number of migrants per household is provided in Table 3. Of the 1,234 households that had returnee migrants, 1,158 (94 percent) had one returnee migrant, while 67 and 9 households had two and three 5 returnee migrants, respectively. Similarly, of the 933 households with current migrants, 803 (86 percent) had one current migrant while 111 had two current migrants. When considering either current or returnee migrants, 1,669 (83.5 percent) of the households had one migrant, 266 had two migrants, 65 had three or four migrants. Table 3: Number of migrants per household Number of migrants Current Total (current or Returnee in the household migrants returnee) migrants None 766 1,067 One 1,158 803 1,669 Two 67 111 266 Three 9 18 59 Four 0 1 6 Total 2,000 2,000 2,000 The vast majority of the migrants (97 percent) in our sample were male. The proportion of current and returnee migrants who are male are largely similar. Similarly, the average age of the migrants is 32 years (between 18 and 69 years). For returnee migrants, the average age is 32.7 years whereas for the current migrants it is 31.2. The difference in age is statistically significant. The average age of those who participated in the phone interview is similar—it is 31.6 years and the age ranges from 18 to 52 years. The average age of the migrant when s/he first moved away is similar for both current and returnee migrants. For the overall sample, the average age for first time migration is 25.2 years; for returnee migrants it is 25.1 years; and for the current migrants it is 25.3 years. For the phone interview respondents, their average age at first migration is 24.2 years. The marital status of current migrants and returnee migrants are different. Of the returnee migrants, 88.5 percent are married; the corresponding figure is 77.8 percent for current migrants. This difference is statistically significant. The education level of the migrants is, on average, low. As Table 4 shows, 7.5 percent of all migrants, and 7.0 percent of phone interview migrants have not completed even one year of schooling. Twenty-five percent of all migrants have five or less years of education, while 81.9 percent have 10 years of education or less. Only 3.2 percent have completed at least a bachelor’s level of education. Hardly anyone in our sample had a Technical School Leaving Certificate (TSLC) or a Technical Diploma. On average, the education level of current migrants is slightly higher than that of returnee migrants and the difference is statistically significant. For example, 79.2 percent of current migrants have 10 years or less of education, while the corresponding figure for returnee migrants is 83.9 percent. 6 Table 4: Education level of migrants in household survey and phone interview Household survey Phone interview Highest level of Frequency Percentage Frequency Percentage education None 181 7.5 23 7.0 Grade One 20 0.8 Grade Two 56 2.3 4 1.2 Grade Three 64 2.7 10 3.0 Grade Four 72 3.0 14 4.2 Grade Five 208 8.7 25 7.6 Grade Six 97 4.0 20 6.0 Grade Seven 144 6.0 22 6.7 Grade Eight 281 11.7 45 13.6 Grade Nine 229 9.5 13 3.9 Grade Ten11 613 25.5 114 34.4 Grade Eleven 43 1.8 7 2.1 Grade Twelve 310 12.9 29 8.8 TSLC 3 0.1 1 0.3 Technical Diploma 4 0.2 Bachelors 73 3.0 3 0.9 Masters and above 4 0.2 1 0.3 Total 2,402 100.0 331 100.0 We present next the figures on the countries current migrants are working. The top four countries where current migrants are working are Qatar (28.5 percent), Saudi Arabia (23.7 percent), Malaysia (22.4 percent) and United Arab Emirates (15.4 percent). The percentage of migrants in other countries is in low single digits. The top four countries in the phone interview are also the same—Qatar (30.5 percent), Saudi Arabia (29.9 percent), Malaysia (22.1 percent), and United Arab Emirates (10.3 percent). Table 5: Countries current migrants are presently working Household survey Phone interview Country Frequency Percentage Frequency Percentage Qatar 307 28.5 101 30.5 Saudi Arabia 255 23.7 99 29.9 Malaysia 241 22.4 73 22.1 United Arab Emirates 166 15.4 34 10.3 Kuwait 40 3.7 9 2.7 Republic of Korea 17 1.6 1 0.3 Oman 13 1.2 4 1.2 Bahrain 12 1.1 5 1.5 Japan 8 0.7 1 0.3 Others 18 1.7 4 1.2 Total 1,077 100.0 331 100.0 11 In the phone interview sample, 15.41 percent had not passed the School Leaving Certificate (SLC) examination; only 19.03 percent had passed SLC examination. This disaggregated figure is not available for the household survey. 7 The migrant profiles from the nationally representative Labor Force Survey conducted in 2017/18 are also largely similar. For example, the top four destination countries are the same as in this survey. As in the current survey, only 2.6 percent of the migrants are female in that survey. Similarly, the average age of migrants outside India as well as in the top four destination countries is slightly over 28 years, and 4 percent of migrants outside India and 5 percent in the top four destination countries aged 18 and above were illiterate. A quarter of absentee migrants in the top four destination countries have five or less years or education, and 81 percent have 10 or less years of education.12 As mentioned above, basic profiles of all migrants in the selected 2,000 households were collected. To obtain information on aspects such as sources of information for job placement abroad, migration experiences, details on current job, participation in vocational or skills training courses and the skills the migrant learned informally, traditionally, or as an apprentice, and further training needs of the migrant workers, one person from each household was identified to talk about these aspects. The details on how this selection was done are explained in Section 2. Of the 2,000 migrants we asked about with the respondent, 1,222 (61.1 percent) are returnees, 109 (5.5 percent) are current migrants temporarily in Nepal, and 669 (33.5 percent) are current migrants who were abroad at the time of the household survey. There are many households in Nepal that have a sizeable number of returnee migrants. Almost 95 percent of these returnee migrants in the sample talked about their migration experiences themselves. The vast majority of the migrants (87.6 percent) are Hindus, while the rest are Buddhist (5.5 percent), Muslim (3.6 percent), Kirat (2.3 percent), and Christian (1.1 percent). Most of the migrants in our sample are Madheshis (33.1 percent including 4.6 percent Tharus), Janajati (23.6 percent), Chhetri (15.6 percent), Brahmin (15.1 percent), and Dalits (11.2 percent). Others comprise 1.5 percent of the sample. On average, the current ever-married migrants have 2.82 children, while the returnees have 2.86 children. There is no statistically significant difference in the number of children between ever-married current and returnee migrants. Number of countries migrants have worked The respondents were asked to state the number of times and countries the migrant had gone for foreign employment outside India. Their responses, given in Table 6, show that a substantial proportion of migrants had traveled for work on multiple occasions and to more than one country. According to the household survey, 61 percent of the migrants had been outside Nepal and India only once, 28.5 percent two times, and 11 percent three or five times. Note that migrants who returned to Nepal for vacation are not counted in their migration episodes. Though almost 40 percent had travelled abroad outside India for work on multiple occasions, only slightly over 31 percent had been to more than one country for work. This means almost 9 percent of the migrants who had traveled abroad on multiple occasions had gone to only one country for work. 12 This is based on the data collected by Nepal’s Central Bureau of Statistics for Nepal Labor Force Survey, 2017/18. 8 Table 6: Migration episodes and number of countries worked by the migrant Number of episodes Number of countries Number Frequency Percentage Frequency Percentage 1 1,211 60.6 1,370 68.5 2 569 28.5 512 25.6 3 178 8.9 110 5.5 4 34 1.7 8 0.4 5 8 0.4 Total 2,000 100.0 2,000 100.0 We consider next the countries worked in by the migrants in our sample. Each migration episode is counted separately—if they traveled to the same country more than once, the country is also counted more than once. Countries worked in by the migrants are provided in Table 7. Malaysia tops the list with 30 percent of the migrants working there. Qatar and Saudi Arabia are the second and third most popular destinations for the migrants, respectively. The top four countries are still the same as in Table 5 for current migrants, but the ordering is different. Table 7: Countries worked by the migrants Country Frequency Percentage Malaysia 918 30.0 Qatar 816 26.7 Saudi Arabia 736 24.1 United Arab Emirates 351 11.5 Kuwait 96 3.14 Republic of Korea 26 0.9 Bahrain 28 0.9 Oman 33 1.1 Others 55 1.8 Total episodes 3,059 100.0 Sources of information on job placement abroad We discuss next the sources of information on job placement. When asked, how the migrants knew about this job opportunity in the last trip, recruiting/manpower agency, agent, friend or acquaintance, and relative were the top four sources of information for them (Table 8). The majority of migrants had used agents and recruiting/manpower agencies. In fact, 86.9 percent of migrants knew about the current job opportunity from a manpower agency or agent. Though relatives and friends/acquaintances were also sources of information in 26.3 percent and 34.1 percent of cases, respectively, only 11.5 percent of the migrants obtained information from relatives and/or friends without help from any recruiting agency or agent. Previous or current employers, radio, social media, and personal contact with employers as sources of information were also low. 9 Table 8: Sources of information on job opportunity in last trip Sources of information Percentage Recruiting/manpower agency 59.9 Agent 57.3 Friend or acquaintance 34.1 Immediate family member/relative 26.3 Radio advertisement 2.35 Previous employer 2.3 Newspaper 1.5 Other media advertisement: social media, Facebook 0.85 Personal contact with the employer 0.5 Current employer 0.3 Since the source of information for a job opportunity abroad and the channel used to migrate can be different, we asked whom or what channel the migrant used to migrate. The responses reveal interesting patterns (Table 9). In total, 94.2 percent of the migrants had used either an agent or a recruiting agency as a channel to migrate, as opposed to 86.9 percent who had used it as sources of information.13 This suggests many more migrants relied on agents and recruiting agencies to migrate for work. Migrants using friends and relatives as channels to migrate comprised 10.5 percent and 12.0 percent, respectively. Moreover, only 3.5 percent migrants had used relatives or friends and not an agency or an agent. The responses from the phone interviews with 331 current migrants are also qualitatively similar—29.8 percent of them had used agents and 69.8 percent had used a manpower agency (85.8 percent had used at least one of these two sources). This goes on to signify the immense role agents and manpower agencies play in sending migrants abroad for work. Table 9: Channels used for migration Household Phone survey interview Channel Percentage Recruiting/manpower agency 76.1 69.8 Agent 41.7 29.0 Immediate family member/relative 12.0 14.8 Friend or acquaintance 10.5 5.4 Previous employer 1.8 1.5 Current employer 0.7 0.3 Employment Permit System 1.0 Migrants’ main job This study also sought to understand the migrants’ main job in their country of destination. The top 12 jobs mentioned by the respondents in the household survey and phone interviews are given in Table 10. The 13 Agent was used by 41.7 percent of migrants and manpower agency by 76.1 percent. Since some had used both agent and manpower agency, the proportion of migrants using either agent or manpower agency is 94.2 percent. 10 jobs varied greatly—for example, 12.9 percent of the respondents in the household survey said laborer, followed by 11.7 percent factory worker, and 8.8 percent as security guard. Table 10: Main job in the country of destination Household survey Phone interview Job Frequency Percentage Frequency Percentage Laborer (including farm laborer) 258 12.9 70 21.2 Factory worker 233 11.7 16 4.8 Security guard 175 8.8 30 9.1 Construction worker 170 8.5 29 8.8 Driver 166 8.3 36 10.9 Electrician 106 5.3 11 3.3 Machine operator 98 4.9 9 2.7 Storekeeper 87 4.4 10 3.0 Waiter and kitchen helper 76 3.8 12 3.6 Salesman 73 3.7 10 3.0 Cook 66 3.3 11 3.3 Domestic worker 54 2.7 4 1.2 One could argue that the respondent in the household survey who is not a migrant may not know the exact work the current migrant does abroad. Therefore, the same question was also asked to the current migrants during phone interviews. Among the jobs mentioned by phone interview respondents, the top ten jobs were the same. Only domestic worker was not in the top 12. As in the household survey, most of the respondents of the phone interviews (21.2 percent) were laborers. The second most prevalent job was driver (10.9 percent), followed by security guard (9.1 percent), and construction worker (8.8 percent). Even when only the main tasks by the returnee migrants are considered, the top 12 main jobs are largely similar—the order of most prevalent jobs changed only slightly (Table 11). Taken together, this goes on to suggest that there is no discernable difference in the jobs performed by current and returnee migrants. Table 11: Main job in the last assignment abroad for returnee migrants Job Frequency Percentage Factory worker 151 12.36 Laborer (including farm laborer) 145 11.86 Construction worker 116 9.49 Security guard 100 8.18 Driver 84 6.87 Machine operator 76 6.22 Electrician 68 5.56 Waiter and kitchen helper 60 4.91 Store keeper 56 4.58 Salesman 44 3.6 Furniture maker 35 2.86 Domestic worker 34 2.78 11 Respondents were also asked about the different tasks in the migrant’s current job or the last job abroad in the case of returnee migrants (Table 12). Though it is difficult to ascertain whether it is a low-skilled or a high-skilled just by looking at one task, it is possible to argue that most tasks are low-skilled based on the responses for different tasks. For example, 53 percent of all migrants lift heavy loads in their job. Only 6.7 percent use computers at work, and 10.9 percent supervise others. Furthermore, less than 5 percent read documents at work. Taking into account the low education attainment by the migrants, 47.7 percent saying they had never worked in Nepal before migrating outside Nepal, and the responses to different tasks in the current job, it is reasonable to argue that most jobs performed by Nepali migrants are low skilled. Table 12: Tasks required in the migrants’ job (in %) Tasks All Returnee Lift heavy loads 53.0 54.4 Drive a vehicle 13.1 12.8 Follow written instructions 36.4 37.2 Follow oral instructions (not written) 81.1 83.3 Use a calculator or do simple calculations 17.3 16.9 Read documents 4.3 4.6 Interact with clients 16.9 17.9 Use computers at work 6.7 6.8 Supervise others 10.9 11.8 Time spent at work abroad When considering the number of hours worked each day by the migrants abroad, most of them worked either 8, 10, or 12 hours a day (Table 13). The mode is 12 hours. On average, migrants in the household survey work 11.1 hours a day. For returnee migrants, it is 11.3 hours a day in the last year they worked outside. For current migrants in the household survey, the respondents who are mostly migrants’ spouses or parents, said it is 10.8 hours a day for the migrants. In the phone interview, the current migrants said they worked, on average, 11.3 hours a day. Table 13: Hours worked by the migrants abroad All migrants Returnee Phone interview Hours Frequency Percentage Frequency Percentage Frequency Percentage 7 or less 9 0.45 7 0.57 3 0.9 8 336 17.03 158 13.01 80 24.2 9 41 2.08 25 2.06 3 0.9 10 404 20.48 262 21.58 65 19.6 11 88 4.46 64 5.27 3 0.9 12 874 44.3 537 44.23 135 40.8 13 30 1.52 23 1.89 14 66 3.35 43 3.54 2 0.6 15 42 2.13 33 2.72 16 59 2.99 46 3.79 17 5 0.25 3 0.25 12 18 or 19 0.96 13 1.06 more 40 12.1 Total 1,973 100.00 1,214 100.00 331 100.0 Regarding number of days a current migrant worked in a normal month or in the case of returnee migrants, the number of days they worked on average in a month the last year they worked outside the country, majority of them had only one day off in a week (Table 14). For example, 68.4 percent of all migrants, 66.2 percent of returnee migrants, and 55.9 percent of current migrants in the phone interview worked 26 days a month. The number of people working all 30 days a month is also significant—17.8 percent in the case of all migrants, 20.4 percent for returnee migrants, and 32.6 percent for current migrants who were part of phone interviews. On average, all migrants worked 26.7 days a month, returnee migrants worked 26.9 days a month, and current migrants worked 27.3 days a month. Table 14: Number of days worked in a month by the migrants abroad All migrants Returnee Phone interview Days Frequency Percentage Frequency Percentage Frequency Percentage worked 22 or less 45 2.3 29 2.4 7 2.1 23–25 34 1.7 3 0.3 10 3.0 26 1,351 68.4 804 66.2 185 55.9 27 32 1.6 15 1.2 28 122 6.2 74 6.1 18 5.4 29 41 2.1 28 2.3 3 0.9 30 351 17.8 247 20.4 108 32.6 Total 1,976 100.0 1,214 100.0 331 100.0 There appears to be a desire on the part of many current migrants to return to Nepal. During the phone interview, when the current migrants were asked whether they would want to continue working in the present country or go elsewhere for work in the next one year, one-third (33.5 percent) said they would want to return to Nepal; 61.6 percent said they would want to continue working in the present country, while 4.8 percent preferred to work in a different country. Migrants’ earnings abroad Questions related to migrants’ current earning (or in the last 12 months while abroad in case of returnee migrants), earning at the start of work in the present company, amount currently saved per month, and the total amount the migrant had sent or brought to the household in Nepal in the past 12 months, including the value of any goods or gifts were also asked. Their responses are summarized in Table 15. 13 Table 15: Monthly earning, savings, and amount remitted by the migrants (NPR)14 All migrants Returnee migrants Variable Obs Mean Std. Dev. Obs Mean Std. Dev. Current monthly earning in a normal month 1,941 46,055.1 26,260.4 1,209 45,018.6 21,716.8 Monthly earning at the beginning 1,937 33,442.6 20,676.2 1,208 31,571.7 14,407.1 Amount per month saved currently 1,849 35,223.7 36,685.2 1,193 33,446.8 18,190.4 Remittance in the last 12 months 2,000 296,165.2 208,975.5 1,222 297,059.4 190,099.3 The average monthly earning for all migrants was about NPR 46,000, as reported in the household survey. This figure includes overtime earnings as well. When they started work in the same company, the average salary was NPR 33,443. Compared to what they were earning in the beginning, the salary had increased, on average, by about 30 percent. The proportion of earning that was saved is also substantial—on average, 76.5 percent of the current earnings were saved by the migrants. This amount is also consistent with the remittance amount they have sent home in the last 12 months. On average, almost NPR 300,000 was sent home by each migrant in the last 12 months, which is 70 percent of the migrants’ yearly savings. These figures are similar when one considers only returnee migrants, further reinforcing the views that household members residing in Nepal also largely knew the earning and saving patterns of the migrants abroad. In the phone interviews which were conducted after the household survey, the earnings of the current migrants both with and without overtime were collected. On average, the current migrants earned NPR 49,155 a month with overtime. The corresponding figure for earnings without overtime is NPR 39,005. The additional earning as a result of working overtime was, on average, NPR 10,150 a month which is slightly over 25 percent of their average base salary. The relationship between earnings and various migrant characteristics were explored using multiple regression analysis. The regression estimates are provided in Appendix A3. Other things remaining constant, earnings for female migrants are substantially lower (33 percent) than for male migrants. Similarly, the earning of those with training is 8 percent higher than for those without training. Migrants with higher level of education also earn substantially more, especially when they have 11 or more years of education. Earnings in some countries are higher than in others. Training and skill needs In this subsection, we focus on the trainings the migrants had taken and the training and skills enhancement they would like to acquire. Questions related to training and skills were asked only to the respondents in the household survey who were migrants (sample size 1,258), and not the migrants’ spouses or other family members as they may not know much about these topics. A subset of these questions was also asked during the phone interview with the current migrants. When the migrant was asked whether s/has ever participated in a vocational or skills training course of any length, free or paid, certificate or not), only 17 percent of them provided an affirmative response. This figure 14 The exchange rate was US$ 1 = NPR 106.36 on May 1, 2018 (https://www.nrb.org.np/forex). The per capita GDP Nepal (in current US$) for Nepal was US$1,039 in 2018 (https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=NP) 14 excludes pre-departure orientation training with little skills content that is provided by the DoFE, which was attended by 54 percent of the migrants even though it is a mandatory training. These figures were largely similar for the responses during the phone interview—60.1 percent said that they have participated in DoFE’s pre-departure orientation training, while only 16.3 percent had participated in other skills training apart from DoFE’s orientation training. Of the migrants who had taken vocational or skills training courses, almost 93 percent of the migrants in the household survey and 91 percent in the phone interview had taken only one such training (Table 16). Of the remaining, 6 percent in the household survey and 9 percent in the phone interview had taken two such trainings. Only 0.9 percent in the household survey had taken three trainings. Table 16: Number of skills training by the migrants Household survey Phone interview Number of Frequency Percentage Frequency Percentage trainings 1 297 92.8 49 90.7 2 20 6.3 5 9.3 3 3 0.9 Total 320 100.0 54 100.0 The major trades in which the migrants have taken training are given in Table 17. In the household survey, the two most common trainings were to become a driver (27.8 percent) and electrician (17.5 percent), followed by cook (9.1 percent) and security guard (8.4 percent). For the current migrants in the phone survey, training to become an electrician was taken by 31.5 percent of those who had participated in training, followed by training to become a security guard and a driver, each with 9.3 percent. Table 17: Trades in which migrants have taken training Household survey Phone interview Trade Frequency Percentage Frequency Percentage Driver 89 27.8 5 9.3 Electrician 56 17.5 17 31.5 Cook 29 9.1 4 7.4 Security guard 27 8.4 5 9.3 Computer application 20 6.3 2 3.7 Waiter 16 5.0 2 3.7 Plumber 12 3.8 2 3.7 Welder 9 2.8 1 1.9 Tailor (Dress maker) 9 2.8 Scaffolding 6 1.9 4 7.4 Others 47 14.7 12 22.2 Total 320 100.0 54 100.0 The duration of these trainings vary greatly. Most of these trainings are for either seven days, fifteen days, one month, three months, or six months. The details on the duration of training are provided in Table 18. The average length of the training was 3.4 months for the migrants in the household survey. 15 Table 18: Duration of the trainings in which migrants participated Training Household survey Phone interview duration Frequency Percentage Frequency Percentage 15 days or less 50 15.6 18 33.3 16 days to 1 month 68 21.3 9 16.7 1.5 to 3 months 120 37.5 16 29.6 4 to 6 months 63 19.7 5 9.3 7 months or more 19 5.9 6 11.1 In the household survey, of those who had taken the training, 89 percent had passed the skill test, while 8 percent had failed it; 3 percent did not know the status of their skill test. Similarly, 79 percent had received the certificate for their training. For 46.7 percent of the migrants who had received a certificate, the certificate issuing body was the NSTB. Almost one-third (31.6 percent) did not know their certificate issuing body. In the phone interview, only 63 percent (34 migrants) of those who had taken training had taken the skill test, and except one, all had passed the test. Of those who had passed the skill test, 28 (84.8 percent) had received the certificate and in 34.8 percent of the cases, the issuing body was NSTB. During the phone interview, migrants currently abroad were asked whether their employers valued the skills test certificate given by Nepali government agencies. A majority (56.8 percent) of the respondents gave an affirmative response. It appears that the vast majority of the migrants who had taken the training were satisfied with the training. In the household survey, 32 percent of the migrants were ‘very satisfied’ with the training, while another 62 percent were ‘satisfied’ with the training. The corresponding figures were 17 percent and 74 percent, respectively, in the phone interview. Only 3 percent in the case of the household survey and 2 percent in the phone interview were ‘not satisfied’ with the training. Similarly, migrants found the training to be largely useful. About 38 percent in the household survey and 22 percent in the phone interview found the training to be ‘very useful’, while 55 percent in the household survey and 59 percent in the phone interview found the training either ‘useful’ or ‘somewhat useful’. Only 5 percent said the training was ‘not very useful’. Skills learned informally, traditionally, or as an apprentice In the household survey, migrants were also asked whether they had learned any skills informally, traditionally, or as an apprentice. Informal skills during a migration episode or anywhere including at home or even in India were also counted. Among the migrants themselves, 45 percent (561) of them had learned skills informally, traditionally, or as an apprentice. Of them, 84.3 percent said they had learned one skill, while 13.4 percent learned two skills and 2.3 percent learned three skills. 16 Table 19: Top five skills acquired informally Skills Frequency Percentage Driver 158 20.9 Construction brick/stone layer mason 63 8.3 Electrician 57 7.5 Furniture maker 55 7.3 Cook 47 6.2 The top five skills they acquired informally are presented in Table 19. These skills are largely similar to the job tasks they are/were doing as a migrant. However, they say most of these skills were acquired domestically. For instance, 53.3 percent say it was acquired domestically, while 5.7 percent say it was acquired in India, and the remaining 41.0 percent say it was in a country outside Nepal and India. When migrants who had acquired skills informally or formally without any certification were asked whether they would like to take a skills test for certification of their skills, 40.1 percent (275) of them gave an affirmative answer. Of those who are interested in taking skills test, 81.8 percent would like to be tested in one skill, while almost all of the remaining wanted to be tested in two skills. Paying for these skill tests does not appear to be a primary concern for many of the migrants, as 72.3 percent of those interested in applying for certification would also be willing to pay for it. Perceived value in training and suggestive evidence based on current data When the migrants were asked how they thought further formal skills training would improve the migrant’s employment prospects, 80.5 percent said it would help the migrant get a better job in the country of destination, a similar percentage (80 percent) also said it would help the migrant get an increase in the salary. In total, 92.7 percent of the migrants thought further formal skills training would improve the migrant’s employment prospects or increase their salary. Only 6 percent said skills training does not make a difference. The migrants in the household survey, on average, think that salary will increase by more than NPR 9,800 per month with additional skills training for the same level of work in their country of destination. Since the average current salary for these migrants is slightly over NPR 45,000, they associate skills training to increase the earnings by more than 20 percent, which is a substantial amount. The current migrants in the phone interview saw even more value in the training. On average, they thought skills training would increase salary by more than NPR 12,200, a premium of 24.8 percent of total earning including baseline, and 31.3 percent of the earning without overtime. The earning data, disaggregated by training status, also provides suggestive evidence that there is a premium to taking training. For example, the current monthly earning in a normal month for those without training is NPR 44,022, while it is NPR 54,350 for those with training—a 23 percent increase (Table 20). Similar difference in monthly earning at the beginning, amount saved per month, and remittance in the last 12 months can be seen by training status. The monthly earning by training status for different occupations is provided in Appendix A1. 17 Table 20: Monthly earning, savings, and amount remitted by the migrants for all migrants by training status (NPR) Without training With training Difference (with Variable training - Obs Mean Std. Dev. Obs Mean Std. Dev. without training) Current monthly earning in a normal 1,583 44,022.33 22,208.22 312 54,350.88 28,379.43 10,328.55*** month Monthly earning at 1,581 32,351.95 20,470.86 313 39,159.28 21,476.94 6,807.33*** the beginning Amount per month 1,530 33,487.64 33,264.72 280 41,560.12 25,916.81 8,072.48*** saved currently Remittance in the 1,606 286,196.10 18,5812.1 320 348,895.3 284,276.50 62,699.20*** last 12 months Note: ***: difference significant at the 1 percent level We also explore whether those who have taken training in the past are also doing work in the same occupation. In the top four occupations people have taken training in the past (cook, security guard, driver, and electrician), almost 62 percent are still working (or last worked) in that occupation. The wage difference between those who are working in the occupation they were trained in and those who are working in a different occupation than what they were trained in is 27 percent (results not shown). The relationship between training and various migrant characteristics were also analyzed using probit analysis. The probit estimates are provided in Appendix A4. Other things remaining equal, migrants with higher level of education were more likely to have participated in the training. Migrants working in Malaysia are less likely to have taken training. Migrants from some districts are more likely than others to have taken training. The gender and being a returnee migrant had no statistically significant relationship with training status. Skills training needs Migrants in the household survey were asked which trades they would like to be trained in if they were to receive further training, and their responses are provided in Table 21. Table 21: Trades migrants would want to be trained in Trades Percentage Driver 28 Electrician 26 Plumber 14 Cook 13 Computer application 9 Machine operator 8 Salesman 7 Security guard 7 Welder 7 18 Trades Percentage Furniture maker 5 Mobile phone repairer 5 A significant portion of migrants wanted to be trained to be a driver (28 percent) or electrician (26 percent), followed by plumber (14 percent) or cook (13 percent). For the phone interview (details not shown), the top three trainings current migrants wanted were to be a driver (19.3 percent), electrician (11.5 percent), and cook (7.9 percent). Other options were suggested by less than 5 percent of the migrants. Only 19 percent of the current and returnee migrants who were the respondents in the household survey did not want further training. In the case of phone interview with the current migrants, 38 percent did not want further training. When asked what the length of these trainings should be, 43 percent in household survey said these training should be for three months, and 23 percent said for six months (Table 22). In the phone interview, 38 percent wanted it for three months, 20 percent for six months, and 19 percent for one month. Table 22: Length of training requested by migrants Length of Household survey Phone interview training Frequency Percentage Frequency Percentage Less than a month 29 2.84 18 9.0 One month 124 12.14 38 19.0 Two months 192 18.81 20 10.0 Three months 438 42.90 75 37.5 Six months 234 22.92 40 20.0 Twelve months 4 0.39 8 4.0 Two years 1 0.5 Total 1,021 100.00 200 It is possible that the training migrants want to take now and what they should have taken before migrating could be different. Therefore, they were also asked which training they think they should have taken before going abroad for work the last time, and their responses are given in Table 23. The top three trades migrants should have been trained, in both the household survey and phone interview, are to be a driver, electrician, and cook. The top three trades are the same trades that migrants have been training in (details in Table 17). Table 23: Top ten trades migrants should have been trained in before migrating Household survey Phone interview Trade Frequency Percentage Frequency Percentage Driver 367 29.20 93 28.1 Electrician 257 20.40 51 15.4 Cook 145 11.50 34 10.3 Plumber 135 10.70 15 4.5 Computer 122 9.70 application 24 7.3 Machine operator 121 9.60 8 2.4 Security guard 95 7.60 13 3.9 Welder 82 6.50 10 3.0 19 Household survey Phone interview Trade Frequency Percentage Frequency Percentage Salesman 71 5.60 4 1.2 Furniture maker 68 5.40 15 4.5 It is possible migrants may perceive their specific current training needs and the needs of migrants in general to be different. The migrants were asked to mention the top three trainings in their view that would be most useful for youth migrants. Table 24 shows the top ten trades respondents said should be provided to the migrants. Note that each person was allowed to mention up to three trades where the GoN could offer skills training and certification to migrants. Fifty percent of the respondents said training and certification of driving should be offered to migrants. Training to be a driver was mentioned by 43.7 percent of the migrants, while training to become a cook, security guard, or plumber were mentioned by more than 20 percent of the migrants in the household survey. Training on being a driver, electrician, cook, or plumber are also listed in the top five trades they would have wanted to be trained now or wished to have been trained before going abroad for work. Table 24: Top ten trades migrants should be trained in Trade Frequency Percentage Driver 1,006 50.3 Electrician 873 43.7 Cook 517 25.9 Security guard 436 21.8 Plumber 399 20.0 Computer application 277 13.9 Machine operator 242 12.1 Salesman 202 10.1 Waiter 183 9.2 Welder 176 8.8 Migrants believe the length of the training they are suggesting for youth migrants should not be very long in duration (Table 25). For example, less than 1 percent say that the training length should be one year or more. Most of them prefer the duration of the training to be either three months (41.5 percent) or six months (28 percent). Table 25: Proposed length of training for youth migrants Length of training Frequency Percentage Less than a month 30 1.5 One month 197 10.1 Two months 351 18.0 Three months 808 41.5 Six months 546 28.0 Twelve month or more 15 0.8 We also asked the migrants about the different ways of reaching out to migrants for offering them training and certification. The responses of migrants who were based largely in Nepal (household survey) and 20 outside Nepal (phone interview) were vastly different (Table 26). According to our respondents in the household survey, the top three ways of reaching out to migrants are media such as radio/television (93.7 percent), newspapers (65.4 percent), and social media (58.7 percent). For the phone interview respondents, the use of social media was the top option with 86.1 percent mentioning it, followed by SMS/Viber/Whatsapp (56.5 percent), and other media such as radio/television and information at the Labor Permits Office with 29.9 percent each. Their responses may also be an indication of the prominent role social media and messaging apps are playing in connecting migrants with others. Table 26: Various options to reach out to migrants Household survey Phone interview Options Frequency Percentage Frequency Percentage Newspapers 1,307 65.4 60 18.1 Other media such as radio/television 1,874 93.7 99 29.9 SMS/Viber/Whatsapp 571 28.6 187 56.5 Information available at post offices, banks, airports 100 5.0 11 3.3 Information available at the Labor Permits Office 155 7.8 99 29.9 Social media 1,174 58.7 285 86.1 Respondents in the household survey were asked how migrants can be supported apart from training, certification, and information, and their responses are given in Table 27. Support in learning languages, access to credit, and specialized trainings were cited by the majority of respondents. Table 27: Services that can be provided to support migrants Service provision Frequency Percentage Support in learning languages 1,129 56.5 Access to credit 1,092 54.6 Specialized trainings 964 51.8 Awareness/orientation programs 788 39.4 Support in insurance and facilitation fee 466 23.3 Linking and facilitating with the foreign employers 465 23.3 4. Further Discussion and Concluding Thoughts This study tries to understand the profiles of the migrants, their work history, skills acquired, and their need for further skills training. The study sample had a good representation of both current and returnee migrants, with the latter representing a majority of the migrant population. The representation of current migrants was ensured by calling them in their destination countries to collect information for the study. The responses to most of the questions are similar in phone interview with current migrants and the household survey with largely returnee migrants and close relatives of current migrants giving confidence that the study is not affected by respondent bias negatively. The studies main findings show that the education level of the migrants is, on average, low. For example, more than 7 percent of the migrants have not completed even one year of schooling, while 80 percent have 10 years of education or less. Current migrants are, by and large, doing tasks that can be perceived as low 21 skilled. Furthermore, a very small fraction of the current and returnee migrants has ever taken vocational training before. Most of the migrants are interested in further skills training—only 19 percent of the migrants in the household survey and 38 percent in the phone interview survey did not want further training. Training on being a driver, electrician, cook, security guard, or plumber are the top five trades migrants wanted to be trained in now or wished to have been trained in before going abroad for work. These jobs are also the major ones they wanted future migrants to be trained in. When asked what the length of these trainings should be, most of them wanted the training to be for three months, followed by six months. Despite not acquiring skills training themselves, there seems to be a perception among migrants that skills training can improve the market value of their labor and certificate can help signal their skill levels to employers. According to the survey respondents, total earning would increase by at least 20 percent (and about 25 percent for phone interview respondents) with additional skills training for the same level of work in their country of destination. There also appears to be a vacuum in testing and certification of skills which could enable returnee migrants to signal their higher skills to employers and facilitate further formal upgrading of their skills in the future. Among the migrants, 45 percent had also learned skills informally, traditionally, or as an apprentice. When these migrants without any certification were asked whether they would like to take a skills test for certification of their skills, 40.1 percent of them gave an affirmative answer. More than 70 percent of them were also willing to pay for the certification. Furthermore, majority of the current migrants in the phone interview survey said that employers outside Nepal and outside India value the skills test certificate given by Nepali government agencies. When asked about different ways of reaching out to migrants to offer them training and certification, the responses differed based on where they are currently located. For the household survey, the top three ways of reaching out to migrants were media such as radio/television, newspapers, and social media. For current migrants, social media was the top option followed by SMS/Viber/Whatsapp. Practical questions related to reaching out to migrants to enhance their skills profile can be a challenging task. However, with proper planning, it can be accomplished. For example, the companies in the destination country are required to provide labor demand, which has to be fulfilled in about five months. So, once the demand is identified, an appropriate skills training lasting about three months can be designed and implemented to prepare skilled manpower to fulfill the demand. The participants from some training and recruiting agencies were receptive to this idea in our informal discussions with them. In addition, the most recent Labor Force Survey suggests that almost 44 percent of the returnee migrants are not in the labor force, presumably because of their intention is to re-migrate.15 If so, contact details of these returnee migrants can be collected at the point of entry and the training opportunities for them can be disseminated on a regular basis. COVID-19 Impact and Implications 15 Ministry of Labour, and Employment and Social Security. 2018. “Labour Migration for Employment—A Status Report for Nepal: 2015/2016–2016/2017. Government of Nepal. 22 According to a report by the Nepal Association of Foreign Employment Agencies (NAFEA), nearly 500,000 to 600,000 Nepalese migrant workers, excluding migrants from India, are likely to return home in the near future.16 About a third of these migrant workers are likely to return due to job loss caused by the pandemic-related economic slowdown; the remaining are likely to return due to completion of their contract tenure. Under normal circumstances most of these returnee migrants would have migrated again. However, the prospect of re-migration is highly unlikely in the current COVID-19 context. And it will take a while for the labor migration sector to bounce back to the pre-COVID-19 level. According to an ongoing survey of International Fund for Agricultural Development or IFAD’s Rural Enterprises and Remittances Project Samriddihi, 60 percent of returnee migrants are not willing to re-migrate under the changed circumstances. Most would like to engage in agriculture-related activities and skills training, access to finance and market links being the most desired elements of support for starting their enterprises.17 With potentially large number of migrants returning to Nepal, their reintegration into the domestic economy has become an important agenda for the GoN. Different government initiatives are ongoing to address the issues of migrant workers ranging from repatriation to proper rehabilitation in the domestic economy. Aimed at facilitating integration programs, the Ministry of Labour, Employment and Social Security has developed and distributed forms to the migrant workers through the diplomatic missions in destination countries to record the skills of the migrants.18 In the absence of detailed data on migrant skills, the data collected through this study can be used to provide the critical information on migrant’s skill mix. The data can be used for informing the different reintegration programs that are under preparation. The data can also be used to narrow down and identify skills for upskilling the migrants and for providing testing and certification services to the returnee migrants. 16 https://kathmandupost.com/national/2020/05/21/at-least-500-000-migrant-workers-want-to-return-home-at-the- earliest-says-report 17 https://www.ifad.org/en/web/latest/blog/asset/41969816#:~:text=Migrant%20returnees%20have%20also%20acqui red,equipment%20repairing%2C%20carpentry%20and%20masonry 18 “Impact of COVID-19 on Nepali Migrant Workers: Protecting Migrant Workers during the Health and Economic Crisis- Challenges and Way Forward.� ILO 2020. https://www.ilo.org/wcmsp5/groups/public/---asia/---ro-bangkok/---ilo- kathmandu/documents/briefingnote/wcms_748917.pdf 23 Appendix A1: Monthly Earning for all Migrants by Training Status and Occupation (NPR) Job Without Training With training Occupation Mean Std. Dev. Freq. Mean Std. Dev. Freq. Domestic worker 35,519 11,594 44 25,461 4,159 4 Health services 63,216 4,548 2 37,800 0 1 Construction worker 39,087 18,373 144 40,871 10,613 15 Farm laborer 40,147 19,327 24 76,768 64,330 5 Other laborer 38,457 24,566 203 41,353 13,988 16 Kitchen helper 35,505 9,504 22 57,204 0 1 Waiter 41,858 15,259 38 40,080 20,478 13 Cook 52,920 25,396 34 67,711 36,402 25 Salesman 52,204 24,870 58 58,804 28,654 11 Security guard 47,205 17,056 127 53,445 16,260 37 Driver 55,118 22,446 95 68,418 30,699 59 Welder 44,123 16,942 30 39,196 17,717 5 Supervisor 72,162 51,926 42 50,515 17,257 8 Electrician 39,331 12,214 75 45,479 18,567 26 Storekeeper 45,402 16,097 72 46,954 29,717 11 Machine operator 47,286 30,408 89 38,049 10,691 5 Factory worker 42,481 19,662 198 47,225 26,385 29 Bartender 25,000 0 1 78,750 0 1 Community livestock assistant 44,283 20,561 3 150,000 0 1 Construction mason 36,454 11,215 36 28,602 0 1 Tailor 32,373 9,123 6 31,933 7,044 3 Furniture maker 42,144 11,372 47 17,992 0 1 Garment fabrication 35,910 10,923 8 44,492 0 1 Hand embroidery 43,176 0 1 33,924 0 1 Plumber 39,558 10,506 25 63,341 34,611 7 Mobile phone repairer 83,616 28,378 3 50,000 0 1 Painter 42,967 14,598 23 46,000 0 1 Housekeeping 37,810 8,632 17 78,580 24,550 3 Scaffolding 46,273 13,673 18 52,058 12,146 6 24 Appendix A2: Descriptive Statistics by Various Groups Currently in Nepal Currently abroad Average Average Monthly Monthly Not Hilly Plain Freq. of income Hilly Plain Freq. of income in Total Trained trained Districts Districts migration in NPR. Total Trained Not trained Districts Districts migration NPR. Total 1,222 181 1,031 462 760 1.5 45,018 778 139 575 263 515 1.57 47,766 (61.1%) (14.9%) (85.1%) (37.8%) (62.2%) (38.9%) (19.5%) (80.5%) (33.8%) (66.2%) Male 1,185 175 1,000 438 747 1.5 45,256 760 133 563 255 505 1.57 47,710 (60.9%) (14.9%) (85.1%) (37.0%) (63.0%) (39.1%) (19.1%) (80.9%) (33.6%) (66.5%) Female 37 6 31 24 13 1.46 37,499 18 6 12 8 10 1.61 50,321 (67.3%) (16.2) (83.8) (64.9%) (35.1%) (32.7%) (33.3%) (66.7%) (44.4%) (55.6%) Edu <=Gr 5 317 24 291 103 214 1.5 40,596 179 24 143 53 126 1.53 40,241 (63.9%) (7.6%) (92.4%) (32.5%) (67.5%) (36.1%) (14.4%) (85.6%) (29.6%) (70.4%) Grade 6-10 711 104 602 290 421 1.53 44,363 434 69 327 163 271 1.6 46,180 (62.1%) (14.7%) (85.3%) (40.8%) (59.2%) (37.9%) (17.4%) (82.6%) (37.6%) (62.4%) > Gr10 194 53 138 69 125 1.37 54,760 165 46 105 47 118 1.56 60,599 (54.0%) (27.8%) (72.3%) (35.6%) (64.4%) (46.0%) (30.5%) (69.5%) (28.5%) (71.5%) 25 Appendix A3: OLS Regression Estimates for Monthly Earnings by the Migrant Variables Log earning Migrant is a female −0.334*** Migrant has taken a training 0.0827*** Migrant is a returnee migrant 0.000683 Migrant has 6-10 years of education 0.0658*** Migrant has 11+ years of education 0.246*** Migrant's age is 30+ years 0.0973*** Migrant worked in Qatar 0.0844*** Migrant worked in Saudi Arabia 0.0505** Migrant worked in United Arab Emirates 0.155*** Migrant worked in other countries 0.362*** Migrant is from Morang district 0.0511 Migrant is from Dhanusha district −0.0564* Migrant is from Mahottari district −0.00155 Migrant is from Makwanpur district 0.204*** Migrant is from Chitwan district 0.215*** Migrant is from Nawalparasi district 0.0119 Migrant is from Tanahun district 0.00809 Migrant is from Rupendehi district −0.0924** Migrant is from Dang district 0.0303 Migrant is from Surkhet district 0.0209 Migrant is from Salyan district −0.0517 Migrant is from Kailali district −0.00492 Migrant is from Kanchanpur district 0.0627 Constant 10.41*** Observations 1,895 Adjusted R-squared 0.161 Note: Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 26 Appendix A4: Probit Estimates for Training Status Variables training Migrant is a female −0.126 Migrant is a returnee migrant −0.0395 Migrant has 6-10 years of education 0.199** Migrant has 11+ years of education 0.531*** Migrant's age is 30+ years −0.0177 Migrant worked in Qatar 0.176* Migrant worked in Saudi Arabia 0.210** Migrant worked in United Arab Emirates 0.254** Migrant worked in other countries 0.490*** Migrant is from Morang district −0.356*** Migrant is from Dhanusha district −0.542*** Migrant is from Mahottari district −0.872*** Migrant is from Makwanpur district 0.0720 Migrant is from Chitwan district −0.0700 Migrant is from Nawalparasi district −0.704*** Migrant is from Tanahun district −1.673*** Migrant is from Rupendehi district −0.753*** Migrant is from Dang district −1.114*** Migrant is from Surkhet district −1.265*** Migrant is from Salyan district −1.662*** Migrant is from Kailali district −1.409*** Migrant is from Kanchanpur district −0.846*** Constant −0.832*** Observations 1,926 Note: Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 27