The Roots of Inequality : Estimating Inequality of Opportunity from Regression Trees

Brunori, Paolo ; Hufe, Paul ; Mahler, Daniel Gerszon

World Bank, Washington, DC

关键词: INEQUALITY;    REGRESSION ANALYSIS;    EQUALITY OF OPPORTUNITY;    MACHINE LEARNING;    LIVING CONDITIONS;   

DOI  :  10.1596/1813-9450-8349
RP-ID  :  WPS8349

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

This paper proposes a set of new methodsto estimate inequality of opportunity based on conditionalinference regression trees. It illustrates how these methodsrepresent a substantial improvement over existing empiricalapproaches to measure inequality of opportunity. First, thenew methods minimize the risk of arbitrary and ad hoc modelselection. Second, they provide a standardized way to tradeoff upward and downward biases in inequality of opportunityestimations. Finally, regression trees can be graphicallyrepresented; their structure is immediate to read and easyto understand. This will make the measurement of inequalityof opportunity more easily comprehensible to a largeaudience. These advantages are illustrated by an empiricalapplication based on the 2011 wave of the European UnionStatistics on Income and Living Conditions.