Do Returns to Education Depend on How and Who You Ask?

Serneels, Pieter ; Beegle, Kathleen ; Dillon, Andrew

World Bank, Washington, DC

关键词: returns to education;    survey design;    field experiment;    development;   

DOI  :  10.1596/1813-9450-7747
RP-ID  :  WPS7747

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

Returns to education remain an importantparameter of interest in economic analysis. A largeliterature estimates returns to education in the labormarket, often carefully addressing issues such as selection,into wage employment and in terms of completed schooling.There has been much less exploration of whether estimatedreturns are robust to survey design. Specifically, doreturns to education differ depending on how informationabout wage work is collected? Using a survey experiment inTanzania, this paper investigates whether survey methodsmatter for estimating mincerian returns to education. Theresults show that estimated returns vary by questionnairedesign, but not by whether the information on employment andwages is self-reported or collected by a proxy respondent(another household member). The differences due toquestionnaire type are substantial varying from 6 percentagepoints higher returns to education for the highest educatedmen, to 14 percentage points higher for the least educatedwomen, after allowing for non-linearity and endogeneity inthe estimation of these parameters. These differences are ofsimilar magnitudes as the bias in OLS estimation, whichreceives considerable attention in the literature. Thefindings underline that survey design matters for theestimation of structural parameters, and that care is neededwhen comparing across contexts and over time, in particularwhen data is generated by different surveys.