ABSTRACT
Using a nationally representative dataset extracted from the Jordanian Labor Market Panel Survey (JLMPS) for the two years 2010 and 2016, we apply both the standard Oaxaca–Blinder and quantile decomposition approaches to provide a more comprehensive distributional analysis of the native-immigrant wage differentials in Jordan over the period 2010–2016. By assessing the contribution of a rich set of labor market characteristics to the distributional wage differentials between the two groups and examining the extent to which such differentials reflect marginalization and discrimination against non-natives in Jordan, we find some interesting results that may hold significant policy implications for policymakers and labor market participants. These results reveal an increase in the mean native-immigrant wage gap over time and a relative intensification, throughout the wage distribution, of discrimination against immigrants among middle-wage workers. Compositional differences, primarily in education, between the two groups explain most of the observed wage gap over the period.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The data that support the findings of this study are openly available for public use through the ERF’s Open Access Micro-data Initiative (OAMDI). Researchers can freely access these micro-data through the ERF Data Portal (www.erfdataportal.com) after completing the required registration procedure.
Notes
1 The JLMPSs are part of a series of labor market panel surveys conducted by the Economic Research Forum (ERF) in collaboration with local Statistical Institutes, from which the Jordanian Department of Statistics (DoS), in some Arab countries since 1998. The micro-data from these surveys are available for public use through the ERF’s Open Access Micro-data Initiative (OAMDI). Researchers can freely access these micro-data through the ERF Data Portal (www.erfdataportal.com) after completing the required registration procedures.
2 One limitation of the wage gap decomposition analysis in this study is the exclusion of observations with no earnings in 2010 and 2016. This may bias our results if the sample of workers differs systematically from those who are unemployed. However, as explained in the methodology section, the only method that can simultaneously address the selection bias issue and decompose quantile wage differentials into the contribution of each covariate is computationally intensive and would prevent us from decomposing the endowment effect into its components. For this reason, we follow some existing empirical studies and ignore this problem when conducting the quantile decomposition analysis.
3 Using pooled cross-sectional data provides more precise estimates due to the increased sample size and the minor statistical complications (Wooldridge, Citation2015).
4 Failure to include appropriate exclusion restriction variables in the wage selection equation, using weak exclusion covariates, or employing exclusion variables that are themselves endogenous may yield inconsistent estimates in the wage second stage equation. It's also known that high collinearity between the inverse Mills ratio, derived from the sample selection equation (first-stage equation) and the other predictors in the second-stage equation may occur when the first and second-stage equations share the same vector of predictors. Then, as noted by Puhani (Citation2000), the model may encounter identification and collinearity problems in the absence of suitable exclusion restrictions.
5 The inverse of Mill’s ratio (λ) is given by the ratio of the standard normal cumulative distribution function to the standard normal density function.
6 In the course of this analysis, the term ‘discrimination effect’ refers to workers with similar human capital characteristics or productive endowments being remunerated differently based on their immigrant status. In this context, it is essential to approach the interpretation of this term with caution, as the estimation may not encompass other pertinent factors. The ‘unexplained’ component encapsulates the combined impact of both discrimination and any factors omitted from the analysis. Readers are advised to bear these limitations when interpreting the decomposition findings.
7 A key concept in the unconditional quantile regression.
8 For more technical details on recentered influence functions and unconditional quantile regressions, we refer to Firpo et al. (Citation2009).
9 We acknowledge that with stronger exclusion restrictions and a more significant Mill's ratio, we might obtain statistically significant wage gaps between natives and immigrants when addressing the selection bias issue for the two years. However, due to data limitations, we cannot identify any other set of valid instrumental variables that may affect job participation without exerting any impact on earnings.