Poster #79 - Measuring Medicaid Eligibility under MAGI: Why Annual Income Misclassifies Eligibility?

• In Event: Friday Poster Reception

Abstract

Introduction/Background
Since the Affordable Care Act (ACA) Medicaid expansion in 2014, eligibility for most nonelderly adults has been determined using Modified Adjusted Gross Income (MAGI). Under MAGI, eligibility is based on monthly income and household size. However, much of the empirical literature relies on annual or aggregated income measures that do not align with program rules. This mismatch can lead to misclassification of eligibility status and attenuation of estimated effects. In particular, limited information on household composition and income timing in cross-sectional data may distort both the numerator (income) and denominator (household size) used to calculate the Federal Poverty Line (FPL), potentially obscuring true eligibility thresholds. 

Purpose/Research Question

This paper examines whether commonly used income measures misrepresent Medicaid eligibility and obscure enrollment patterns at the 138 percent Federal Poverty Line (FPL) threshold. It asks whether constructing eligibility based on monthly rules (lowest monthly earnings) reveals discontinuities on Medicaid enrollment. The paper also quantifies the extent to which reliance on annual income introduces attenuation bias in estimating enrollment responses. 

Data
I use panel data from the Survey of Income and Program Participation (SIPP), 2013–2016, which provides detailed monthly information on earnings and household composition. 


Research Design and Methods
I construct a rule-based measure of Medicaid eligibility that follows federal MAGI regulations (42 CFR § 435), incorporating household definition and monthly income. Eligibility is defined using the lowest monthly earnings within a calendar year. I compare this measure to standard approaches based on annual and median monthly income to evaluate how measurement choice affects the presence and magnitude of enrollment discontinuities at the 138 percent FPL threshold. 

Results/Findings

Using lowest monthly earnings, Medicaid enrollment exhibits a 7.8 percentage-point drop at the 138 percent FPL threshold among childless adults. No comparable discontinuity is observed when eligibility is measured using annual or median monthly income. Simulations show that cross-sectional datasets, such as the American Community Survey, attenuate eligibility discontinuities by construction. 


Conclusion/Implications
These findings demonstrate that misalignment between Medicaid eligibility rules and commonly used income measures can obscure key features of program participation. Reliance on annual income attenuates eligibility discontinuities by construction, helping explain inconsistent or null findings in the existing literature. Aligning empirical measurement with monthly eligibility rules is essential for accurately evaluating Medicaid policy. This framework also helps reconcile why some studies identify Medicaid coverage among households with annual incomes well above the eligibility threshold, highlighting the importance of measurement in policy analysis.

Author

• Mingjian Li, University of Illinois, Chicago
  Presenting Author