Publication Type: Paper/Poster Proposal
Abstract: Surveys attempting to measure the number of insured consistently find that
respondents underreport Medicaid coverage. In a study of Minnesota Medicaid recipients, Call, et. al., for example, find a number of reasons for the undercount and conclude that the undercount has a negligible impact on estimates of the uninsured for Minnesota. However, they also note that the study may not generalize to other locations with very different populations. In this study, we consider Medicaid underreporting in Louisiana and its implications for survey estimates of uninsured adults and children. Data for the analysis are based on a random sample of Medicaid recipients in Louisiana, allowing comparisons of reported insurance status for these Medicaid recipients. The paper will also consider the implications of Medicaid underreporting on state-level and national-level estimates (CPS data) of Louisiana's uninsured population. As a final note, with its unique charity hospital system and high poverty level among Louisiana natives, Louisiana provides an interesting and unique context for examining the problem of Medicaid underreporting.

Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: OBJECTIVE: To assess the frequency of differences in race/ethnicity designations between survey responses and Medicaid administrative data.
METHODS: Medicaid enrollees in South Florida were surveyed by telephone using the Consumer Assessment of Health Plans Study (CAHPS) 2.0 Medicaid Adult instrument. The final dataset included 1023 individuals, with 998 giving valid responses to the race/ethnicity survey questions. Race/ethnicity survey responses were compared with Medicaid administrative data for the same individuals.
RESULTS: Overall, about a third (34 percent) of survey respondents gave a different survey response for race/ethnicity than was listed in the Medicaid database. The percentage of survey respondents reporting they were "Black" was virtually identical to the administrative database, with just a few individuals giving responses that varied. However, in all other categories, there were marked differences between the Medicaid recipients' survey report of race/ethnicity and their race/ethnicity in the Medicaid files. The survey responses yielded more Hispanics and fewer Whites than the administrative data. Almost a quarter of survey respondents had been classified as "other" in the administrative data (24.4 percent), but fewer than 1 percent reported "other" in the survey. More than twice as many reported Hispanic origin in the survey than in the administrative records (46.1 percent to 17.8 percent).
DISCUSSION AND CONCLUSIONS: Some differences may result from the mode by which the data were collected; the telephone survey relied entirely on self-report, while the classification in the administrative database may have been made by a third party (e.g., outreach worker) observing the recipient's physical characteristics. These findings are important to survey researchers since administrative records are often used as sampling frames for surveys of program recipients, and the race/ethnicity data listed there provide a basis for sampling design and weighting. Further, administrative data are sometimes considered the "gold standard" against which the quality of survey findings is measured.

Publication Type: Proceeding
Abstract: Challenges in Medicaid implementation present a potential public policy crisis at the intersection of state fiscal policies and health care in a federal system. State budgets have been strained since the national economic downturn in 2001 and remain unstable as state Medicaid expenses continue to rise (Kaiser Commission on Medicaid and the Uninsured 2005). The integrity of the Medicaid program is threatened as Medicaid budget shortfalls have occurred in more than half the states, and are anticipated to occur in 40 states in 2006 (National Association of Budget Officers 2004).
Devolution of responsibility from the federal government continues to present basic questions about state capacity to perform the functions devolved (Kincaid 1998). State revenues are an essential element of the Medicaid service delivery equation. However, it is not clear that state revenue streams can provide a sufficiently stable source of funds in the devolution environment (Tannenwald 2002). Lav, McNichol and Zahradnik (2005) argue that states may exist in structural deficit, with a chronic inability to grow along with growth in the economy and the cost of government (2005: 1). This paper tests the influence of state structural deficits along with other state contextual factors that may explain current state Medicaid budget shortfalls. The influence of state structural deficits and other contextual factors on Medicaid implementation has implications for state capacity to implement policy in the American federal system.

Publication Type: Conference Paper/Unpublished Manuscript
Abstract: This paper is directed at understanding the relationship between state political belief systems, state budgetary resources and state Medicaid enrollment decisions and their effects on the level of state per capita Medicaid hospital expenditure. It tests the theory that there are three different political subcultures that relate state political belief systems to state and local taxation decisions, to the allocation of state taxes to the Medicaid program and their effects on per capita Medicaid hospital expenditure controlling for per capita state personal income and for state Medicaid enrollment effort. It is posited that the independent variables state public political ideology represents a political belief system that is distinguishable from the independent variable state political subculture in its impact on the dependent variable per capita Medicaid hospital expenditure.

Publication Type: Abstract
Abstract: Title: The Effects of Medicaid Enrollment Procedures on the Insurance Coverage of Children

Rationale: In the 1990s states began to change the administrative procedures associated with applying for Medicaid, first easing procedural requirements and subsequently tightening them. While states have held eligibility levels constant in the face of recent budget shortfalls, the same is not true for administrative procedures; eleven states have tightened the enrollment requirements for Medicaid applicants since 2001. Theory predicts that parents with varying preferences regarding insurance coverage and different costs of enrolling will differentially respond to changes in enrollment procedures. Further, parents with lower education levels face more difficulties in navigating Medicaid’s procedural requirements than their peers with higher education levels, suggesting that the insurance take-up of children in families with parents with low education levels is disproportionately affected by enrollment procedures.

Objective: This work will analyze the effects of administrative procedures on the insurance coverage of Medicaid-eligible children. Also, it will test for differential effects between children in families with lower education levels versus those in families with relatively higher education levels.

Data and Methods: The dependent variable of interest is insurance status: whether a child is publicly insured, privately insured, or uninsured. The 2001-2005 rounds of the Annual Social and Economic Supplement of the Current Population Survey (corresponding to the years 2000-2004) will be used for the analysis. The primary independent variable of interest is an index of Medicaid enrollment requirements comprised of: the presence of presumptive eligibility (coverage is provided immediately upon application receipt instead of after the verification and processing of the application); the requirement of an asset test; adoption of a 12 month versus a 6 month renewal period; the requirement of a face-to-face interview at application (in contrast to a mail-in application); and the requirement of applicant income verification (in contrast to the sufficiency of self-reported income). The other major independent variable of interest is maternal education, which is also available in the Current Population Survey.

Regression models for categorical data will be estimated for the main effects of enrollment procedures and maternal education on children’s insurance status as well as the interaction of these two independent variables. Covariates will include: age, sex, race, household income, health status, household composition, urban/rural residence, the state unemployment rate, and parental employment characteristics. We will consider models with both fixed and random effects in order to account for time-invariant unobserved heterogeneity among states that may influence both enrollment procedures and insurance coverage.

Conclusions: States have been actively changing Medicaid enrollment procedures throughout the past decade. Very little research has been done on the effects of these policy changes on children’s insurance coverage. This study’s aim is to provide evidence on the effects that enrollment policies have on the insurance coverage of Medicaid-eligible children and if they exert differential effects on children in families with low levels of maternal education.