The University of Michigan
555 South Forest Street
Third Floor
Ann Arbor, MI 48104-2531

T 734-936-9842
F 734-998-6341



Minority full-time workers are less likely than white full-time workers to have health insurance through their employment. A study by Irena Dushi and Marjorie Honig, funded by the Economic Research Initiative on the Uninsured (ERIU), examines whether this gap can be explained by differences in health insurance offers, take up, or neither.

Trends in employment-based coverage among full time workers across race and ethnic groups.

  • Over 1988-2001, coverage by employment-based health insurance among full time workers fell from 78% to 73% for males and from 71% to 68% for females.
  • The trend varied across racial and ethnic groups. For non-Hispanic black full time workers, there was no statistically significant decline among either men or women. Coverage rates fell for Hispanic and non-Hispanic white men. Among women, the decline was marginally significant for Hispanic women (p<.10), but statistically insignificant for non-Hispanic white women.
  • Viewed as a majority-minority gap, the gap between non-Hispanic black and white male full-time workers remained the same between 1988 and 2001 in percentage point terms. There was no black/white gap among non-Hispanic women, and that did not change over the period. In contrast, the non-Hispanic white/Hispanic gap grew, from 17 to 21 percentage points among men and from 10 to 15 points among women.

Offer and Take Up Differences: Descriptive Statistics

  • There were no statistically significant differences in offer rates between non-Hispanic white and non-Hispanic black men who worked full time. Between non-Hispanic white and non-Hispanic black women, blacks who worked full time had a lower offer rate (62% v 67%).
  • There are larger differences between Hispanics and non-Hispanic whites. Offer rates were lower for Hispanic male (54% v. 70% for non-Hispanic whites) and female (54% v. 67% for non-Hispanic whites) full time workers.
  • There are fewer differences in take up (conditional on offer) between members of majority and minority groups. Across household types, the only statistically significant differences with non-Hispanic whites were Hispanic men in dual earner households (76% v. 86% for non-Hispanic whites) and single Hispanic females (78% v. 87%, p<.10.)
Offer and Take Up Differences: Descriptive Statistics: Multivariate Models
  • In a model of own employer offer and take up among dual earner married households, estimated using spouse's personal and job characteristics as instruments for spouse's coverage and a sample selection term, offer rates for Hispanics are lower than for non-Hispanic whites: 8.7 percentage points lower for Hispanic husbands than non-Hispanic white husbands and 9.7 percentage points lower than for white wives, half to two-thirds the difference in descriptive statistics.
  • In contrast to the descriptive statistics, the regression model does show statistically significant differences between black and white males, with black offer rates 7.3 percentage points lower than for white males.
  • Take up rates showed no racial or ethnic differences.
  • Among other household types, Hispanics also had statistically significantly lower offer rates than non-Hispanic whites in one-earner households, but no difference for single person households.
Decomposing Racial and Ethnic Differences
  • In Blinder-Oaxaca decompositions using linear probability models, job and demographic characteristics fully explained the white-minority gap in take up for working husbands in one earner households and for single men. They could account for 36% of the difference for husbands in dual-earner households, 71% for wives in dual-earner households, and 40% of the difference for single women.

Disparities among full time workers are larger between Hispanics and non-Hispanics than between whites and blacks. These differences stem from differences in offer rates as opposed to differences in take-up rates. Reductions in the coverage gap between racial groups require a focus on employers or patterns of employment instead of take-up.

The paper considers health insurance among full-time workers. The share of non-Hispanic whites, non-Hispanic blacks and Hispanics who work full time is not the same. Part of the effect of being non-white may work through the decision to work and obtaining part time v. full time work.

Although the analysis suggests an important role for employer offer decisions in contributing to the coverage gap, individual decisions are likely important because workers may express their demand for coverage not through take-up, but through the characteristics of their employer, characteristics that include whether to work for an employer that offers or does not offer coverage.

Survey of Income and Program Participation. 1996 Panel.; Wave 5 benefits topical module (interviews from July to October, 1997.) For full time workers, sample size is 2,536 in dual-earner households, 1,687 in one-earner households, and 5,169 single worker households. Also, descriptive statistics from the May 1988 Current Population Survey (CPS) Employee Benefits Survey and the 1995-2001 February CPS Contingent Work Supplements.

Probit models of health insurance offer and take up, estimated separately for three household types: dual earner married households; one-earner households; and single person households. For dual-earner households, spouse's health insurance offer is instrumented with predicted offer.

Offers or Take-Up: Explaining Minorities' Lower Health Insurance Coverage
Irena Dushi, International Longevity Center USA, and Marjorie Honig, Hunter College and CUNY Graduate School.

Conference paper presented at ERIU Vulnerable Populations Research Conference, October 2004

ERIU Working Paper #40 (Adobe PDF)

Back to top

Funded by The Robert Wood Johnson Foundation, ERIU is a five-year program shedding new light on the causes and consequences of lack of coverage, and the crucial role that health insurance plays in shaping the U.S. labor market. The Foundation does not endorse the findings of this or other independent research projects.