Migration decisions and destination choices

Abstract

We used data collected from a face-to-face interview survey of 395 randomly selected farm households in Central Nepal’s Chitwan district to understand the migration decision-making processes and migration destination choices. This district was chosen because it represents the microcosm of whole Nepal due to the presence of migratory population from all over the country within the tropical and sub-tropical climatic districts. We used probit and multinomial logit regression models to discern the individual and family migration decision-making behaviors and to understand what determines the choice of migration destinations. Results suggested that migration decision-making is based on (a) the presence of a large number of young males in the family; (b) having fewer males with secondary education; (c) having more females with secondary education; and (d) higher household wealth. These ‘a–d’factors positively affect migration decision-making. Destination choice is generally dictated by individual characteristics and the economic potential of a destination site. Young, unmarried, male heads of families with a relatively large number of adult males but relatively low land holdings and wealth status choose Malaysia; older people from families with a fewer number of educated females and low land holdings choose India, and relatively less-educated females from families with less-educated heads of households but higher land holdings choose the Gulf Cooperation Council countries as their migration destination.

Keywords:

• Migration decision
• destination choices
• multinomial logit
• Nepal

Notes

1 Please see Chesnais (1992) for more information about the stages of demographic transition.

2 http://www.worldbank.org/en/topic/migrationremittancesdiasporaissues/brief/migration-remittances-data.

3 https://www.theguardian.com/commentisfree/2017/mar/19/qatar-world-cup-slavery-migrant-workers.

4 Nepali labor migrants have received work permits from Nepal’s Ministry of Labor and Employment to work in 153 countries over the period ranging from 2008/09 to 2016/17.

5 https://data.worldbank.org/indicator/BX.TRF.PWKR.CD.DT?locations=NP.

6 In 2017, Qatar, Saudi Arabia and India were the top three countries sending remittance to Nepal. Source: https://www.pewglobal.org/interactives/remittance-flows-by-country/.

7 After the promulgation of a new constitution in 2015, Nepal consists of a total of seven provinces.

8 In 2014, two more municipalities (Khairahani and Chitrawan) were added, making the total municipality number four in the district.

9 Based on the NRCS/USDA definition, the animal unit is 1.0 for a cow, horse or ox (an ox is worth the same as a mature cow under 454 kg with an animal unit of 1.0), 0.7 for a buffalo, 0.1 for a goat (a goat is worth the same as a sheep or a lamb with an animal unit of 0.1), 0.3 for a pig (the animal unit for a pig between 25 and 136 kg is 0.3), 0.033 for a chicken, 0.01 for a duck and 0.003 for a pigeon (a pigeon is worth the same as a chicken under 2.3 kg with an animal unit of 0.003). Source: https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs144p2_051957.pdf

10 The wealth index is based on asset ownership indicators derived using weights that can be obtained through principal component analysis (PCA) (Filmer and Pritchett, 2001; McKenzie and Rapoport, 2007; Smits and Steendijk, 2015; Vyas and Kumaranayake, 2006). The wealth index of a household $i$ can be defined as:

$W_i=\sum _k{\alpha }_k\left(\frac{x_{\mathit{ik}}-{\overline{x}}_k}{{\sigma }_k}\right)\left(18\right)$

Here, $x_{\mathit{ik}}$ is the value of asset $k$ for household$i,$ ${\overline{x}}_k$ is the sample mean, ${\sigma }_k$ is the sample standard deviation and ${\alpha }_k$ represents the weight for each variable $x_{\mathit{ik}}$ for the first principal component that is calculated using PCA. Details about weights for the assets can be obtained from McKenzie (2005). Here, the wealth index can take positive or negative values. As indicated by Mora and Taylor (2006), a positive value of a wealth index represents that the household’s wealth is above the sample wealth average, and a negative value represents that it is below the sample wealth average. We create a wealth index based on McKenzie (2005), McKenzie and Rapoport (2007), and Vyas and Kumaranayake (2006). First, we create all the variables of interest in the form of a dichotomous value (1 = Yes and 0 = No) (Table 2). Then, weights for each asset are generated through PCA using a method described by O’Donnell et al. (2008). We have provided the summary of the principal component correlation for 25 components and a scree plot of eigen values in Table A1 and Figure A1, respectively (appendix). We provide the summary of the first four principal components for all the variables, which contain more than 85% of the variance in Table A2 (appendix). We predict the value of the wealth index for each household based on the first four principle component.

11 There are insufficient observations at the country level for these remaining destinations.

Additional information

Notes on contributors

Madhav Regmi

Madhav Regmi holds a PhD degree in economics with specialization in agricultural economics. He received an outstanding MS student award in 2014 from the Department of Agricultural Economics and Agribusiness at the Louisiana State University. His research interest is in finance, production economics and international development economics. He has recently published papers in journals such as Food Security and Agricultural Finance Review.

Krishna P. Paudel

Krishna P. Paudel is the Gilbert Durbin Endowed professor in the Department of Agricultural Economics and Agribusiness, Louisiana State University (LSU) and LSU Agricultural Center, Baton Rouge, Louisiana, USA. His research work focuses on environmental and resource economics and international development economics. He has published 87 papers in a peer-reviewed journals. He is the co-editor of recently published Routledge Handbook of Agricultural Economics.

Keshav Bhattarai

Keshav Bhattarai is a professor in the School of Geoscience, Physics, and Safety at the University of Central Missouri. He works on GIS (spatial analysis, network analysis, geostatistical analysis and 3-D modeling), remote sensing (land-use and land cover changes), political ecology, migration, cultural landscape, tourism, Asian geography, and forest economics. He has published several peer-reviewed journal articles in geography, tourism and development areas. His recent coauthored published book is titled Historical Dictionary of Nepal.