Who Benefits from Higher Education in Low- and Middle-Income Countries?

Abstract

In this article, we investigate how higher education contributes to the employment and earnings of individuals in labour markets, and whether social origins play a role in the financial benefits from higher education. We focus on these questions in nine low- and middle-income countries: Armenia, Bolivia, Colombia, Georgia, Ghana, Kenya, Laos, Macedonia, and Vietnam. We use the recent Skills Towards Employability and Productivity (STEP) surveys of urban labour force participants to examine individuals’ educational attainment, labour market participation, and earnings. Using logistic regressions, we find that individuals from disadvantaged origins are less likely to obtain a higher education degree. We find that in most of these countries, individuals who have earned a higher education degree are significantly more likely to be in the labour force and find employment, and enjoy sizable earnings premia. The findings are fairly robust with regard to the samples of individuals examined, and the methods used to measure earnings premia. Finally, we find little evidence that the earnings premia from higher education vary by social origins or the likelihood of an individual completing a degree. These results suggest that the benefits from higher education are comparable for individuals from disadvantaged and advantaged social origins.

Acknowledgments

Two anonymous reviewers, Javier Botero, Sangeeta Goyal, Toby Linden, and Francisco Marmolejo provided detailed comments on earlier drafts. Ji Liu provided research support. We are grateful to Sean Kelly and Marc DeFrancis for detailed edits and comments on an earlier draft. We also thank Luis Benveniste, Harry Patrinos, Halsey Rogers, Sebastian Taborda, Namrata Tognatta, and seminar participants at the World Bank for comments and discussions. The findings, interpretations, and conclusions expressed in this study are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. The original data are freely downloadable from the World Bank’s STEP Skills and Measurement Program website (http://microdata.worldbank.org/index.php/catalog/step). The authors are happy to share the data and code with bona fide researchers. Please email the authors for access to the data and code.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary Material

Supplementary Materials are available for this article which can be accessed via the online version of this journal available at https://doi.org/10.1080/00220388.2018.1528351

Notes

1. The economic literature on ‘sorting’ may consider both the economic and sociological explanations (Fernandez, 2001). That is, sorting models consider that ability and motivation affect sorting into higher education by reducing the cost of education, and that social origins affect the preferences and likelihood of sorting into higher education. Sorting models may also consider heterogeneous returns to education.

2. We did not include three STEP participants: Sri Lanka, Ukraine, and Yunnan province in China. Sri Lanka is not included because we would have to consider a region (South Asia) with only one country. Ukraine is not considered because we ran into issues with small shares of male bachelor’s degree holders in Ukraine. The Yunnan province was not included because the complex interplay of Yunnan with other provinces in China is beyond the scope of this article.

3. We construct a parental engagement variable that measures the quality of time spent using the STEP question: ‘When you were attending primary school, did either of your parents/guardians actively keep themselves informed of your exam/test results or grades?’ Response choices include ‘Yes, always or almost always’, ‘Yes, sometimes’, and ‘No, never or almost never’. Previous studies using similar measures of quality of time as an indicator for parental engagement and investment in education include Sandberg and Hofferth (2001). We acknowledge the possibility of reference bias arising because of retrospective reporting. We have compared the distribution of socio-economic status at age 15 to the distribution of current assets as part of our preliminary analysis. Using information on dwelling characteristics and types of assets, factor analysis was used to create an asset index for each of the countries in the sample. Measures of assets and dwellings with extremely skewed distributions, agricultural assets, and those showing low factor loadings were excluded from the asset index. We find no evidence that the retrospective data are upwardly biased.

4. For Ghana, Kenya, and Macedonia, we did not create regional dummy variables because the STEP survey data did not support them.

5. PSM models were implemented using the ‘teffects’ command in Stata 15. We used a nearest neighbour matching (one or more) and a caliper of 0.05. It should be noted that the teffects command for PSM in Stata 15 jointly estimates the propensity scores and the ATE/ATT, and thus does not reduce the sample for common support as is true for the Stata command psmatch2 in earlier versions of the programme.

6. Consistent with the sorting idea, our results show that individuals with more highly-educated parents were themselves more likely to obtain a degree. But despite the sorting, we found no strong evidence that those from lower social origins have different earnings premia than others, so once they have been sorted the labour market gains are the same.

7. The heterogeneous treatment effects were estimated in Stata 15 using the ‘hte sm’ routine developed by Jann, Brand, and Xie (2010).

Additional information

Funding

This study was financially supported by the World Bank as a part of the background papers commissioned for the World Development Report 2018.