![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
This paper quantifies the short to medium term impact of rural-urban migration on an individual’s subjective well-being in South Africa between 2008 and 2012. We work through different econometric specifications; using instrumental variables to control for endogeneity caused by shock-induced self-selection, and Propensity Score Matching to control for migration self-selection bias. We find that rural-urban migration leads to a decrease in subjective well-being by 8.3 per cent. We suspect that the decreased well-being is a result of false expectations and changing relative groups used to peg aspirations, as well as the emotional cost of being away from family and a home environment.
Acknowledgements
The data and code used in the analysis can be provided to bona fide researchers on request at [email protected].
Disclosure statement
No potential conflict of interest was reported by the authors.
ANNEXURE A: Description of Explanatory Variables
Notes
1. This will only be the case if it can be shown that individuals not completing all the necessary NIDS questions used for this analysis have similar characteristics.
2. Diener, Emmons, Larsen, and Griffin (Citation1985) developed a new index to measure global life satisfaction, the Satisfaction with Life Scale (SWLS), Kahneman and Krueger (Citation2006) developed the U-Index, and the Day Reconstruction Method (DRM) was also established (Krueger & Schkade, Citation2008).
3. The decision to model life satisfaction as a continuous variable was further corroborated by applying the same analysis models through an ordered probit (tested for Model I and II) and yielding very little difference in the findings.
4. The fixed effects, and not random effects, model was selected as we are controlling for the variation within individual entities and do not anticipate that the variation across entities should influence the dependent variable (Wooldridge, Citation2006)
5. Other propensity score matching techniques are nearest neighbour, kernel and local linear matching (Khandker et al., Citation2010).
6. Beta is estimated as (coefficient of x/() to enable comparison between coefficients of dummy and continuous variables.