Earned Income of the Economically Disadvantaged Population: The Association with Public Transfers and Other Individual- and Household-Level Factors

Background

The core idea of welfare reform in 1996 rooted in the work ethic that is predominant in the American society and in the public opinion that welfare should be provided to the deserving poor who participate in labor market. This study is about individual earned income and the related factors at different level components. Focusing on the socioeconomically disadvantaged who live in households receiving public assistance, it aims to examine the association between earned income of the disadvantaged population, means-tested transfers, and other individual and household level covariates.

Research Questions

What is the association between means-tested transfers and total earned income at individual-level, controlling for other individual and household level factors? What other factors are associated with total earned income?

Does the effect of total means-tested cash transfer on the total earned income differ by males and females?

What portion of the variance in the total earned income is associated with household types nested in state level and individual-level factors nested in household- and state-levels? Do the within-household effects of total means-tested cash transfers differ from the between-household effects, controlling for the other factors?

Methods

This study uses the Survey Income and Program Participation for the panel 2008, collected by the U.S. Census Bureau. This study focuses on individuals in workforce ages who live in the households receiving public assistance and uses multilevel modeling including individual, household, and state levels. The multilevel design covers models with variance component, random intercept with main effects, interactions and the mean of cash transfers. This study includes discussion on comparison of the constructed models respectively.

Results

A model represents 0.86% of the total variance in total person’s earned income is associated with the state-level, and 4.1% is attributed to the household-level. The models including household type variable and an interaction describe that about 1% of total variance in total person’s earned income is associated with the state-level, and 3.8% is attributed to the household-level. The result represents that per 100 dollar increase in means-tested cash transfer, the estimated mean total person’s earned income decreases by 5.28%, controlling for the other covariates. It also indicates that a female individual’s total earned income is estimated 27.39% lower than total earned income of a male individual, controlling for the other explanatory variables. The coefficient of the interaction estimates that the gender gap between females and males on the total person’s earned income is reduced according to each extra 100 dollars of means-tested cash transfer, controlling for the other covariates (p-value < 0.05).

Conclusions

The within-household effect of total means-tested cash transfer differs from the between household effects, controlling for the other factors. And, study results partly support an existing study on the relationship between employment sector and welfare recipients. This study can address the influences of household-level policy for income support on individual work participation and their earnings. For an effective income policy, policymakers may need to consider the within-household effects of public programs may differ from the between-household effects.