What drives social returns to education? A meta-analysis

•A person’s productivity can benefit from the schooling of the people with whom that person interacts.•We study the drivers of these external effects by reviewing 32 studies that examine these effects.•We find that such schooling spillovers are significant but fall with the economic development of t...

Full description

Saved in:
Bibliographic Details
Published in:World development 2021-12, Vol.148, p.105651, Article 105651
Main Authors: Cui, Ying, Martins, Pedro S.
Format: Article
Language:English
Subjects:
Bias
Convergence
Economic analysis
Economic development
Economic growth
Economics
Education
Education externalities
Externality
Meta-analysis
Returns to education
Spillover effect
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•A person’s productivity can benefit from the schooling of the people with whom that person interacts.•We study the drivers of these external effects by reviewing 32 studies that examine these effects.•We find that such schooling spillovers are significant but fall with the economic development of the country.•On the other hand, spillovers increase when schooling levels are more uneven across the population.•Spillovers are larger when considering co-workers in the same firm (as opposed to the same region or industry). Education can generate important externalities that contribute towards economic growth and convergence. In this paper, we study such externalities and their drivers by conducting the first meta-analysis of the social returns to education literature. We analyse over 1,000 estimates from 32 journal articles published since 1993, covering 15 countries of different levels of development. Our results indicate that: (1) there is publication bias (but not citation bias) in the literature; (2) spillovers slow down with economic development; (3) tertiary schooling and schooling dispersion increase spillovers; and (4) spillovers are smaller under fixed-effects and IV estimators but larger when measured at the firm level.
ISSN:0305-750X
1873-5991
DOI:10.1016/j.worlddev.2021.105651