Search

Growing research indicates that Medicaid expansions reduce Supplemental Security Income (SSI) participation, although the magnitude of effects may vary with the presence of other health policy reforms. We examine how a series of Medicaid expansions for childless adults before and after implementation of the Patient Protection and Affordable Care Act (ACA) impact SSI participation among childless adults. We use the Medical Expenditure Panel Survey to compare the change in SSI participation in states that expanded Medicaid coverage relative to those that did not during two time periods: 2001–2013 and 2001–2015. On average, SSI participation declined by a nonstatistically significant 0.10 percentage points (95% confidence interval [CI] = [−0.54, 0.32]) from a baseline SSI participation rate of 2.2% following Medicaid expansions implemented between 2001 and 2015. When restricted to the pre-ACA era, SSI participation declined by a nonstatistically significant 0.30 percentage points (95% CI = [−0.77, 0.16]). Although the direction of the point estimates suggests that Medicaid expansions before and after implementation of the ACA may be associated with reduced SSI participation, the imprecision of the estimates due in part to the sample size precludes this conclusion. To guide states in decision-making, it is essential to understand how increased Medicaid availability impacts SSI participation under different policy environments.

With increasing drug prices in the past decade, affordability and medication adherence are a growing concern for near-poor older adults, especially for those who are not receiving Low-Income Subsidy in Medicare Part D. SeniorCare is a pharmaceutical assistance program in Wisconsin for near-poor older adults, providing comprehensive prescription coverage with flat copayments.To evaluate five-year trends in financial hardship and medication adherence and to examine factors associated with these outcomes in SeniorCare members.SeniorCare program enrollment and pharmacy claims data from 2014 to 2018 were used. The study population was near-poor older adults in SeniorCare with annual family income ≤200% of the federal poverty level. Financial burden was assessed using the proportion of total annual out-of-pocket costs to total annual income. Medication adherence was assessed by adapting the measures endorsed by the Pharmacy Quality Alliance and National Quality Forum. Descriptive statistics and independent t-tests were used to evaluate the trends, and multivariate logistic regressions were conducted to examine factors associated with financial burden and medication adherence.From 2014 to 2018, mean annual out-of-pocket costs per member declined by 3.7% (p 0.001) for all drugs, while those for specialty drugs increased by 31.2% (p 0.05). Around 3.3% spent more than 5% of their income for prescription drugs in 2014, which decreased to 2.4% in 2018 (p 0.001). The proportions of adherent patients increased from 78.1% to 81.2% (p 0.001) for diabetes medications (excluding insulins), from 77.3% to 79.5% (p 0.001) for statins, and from 79.8% to 80.8% (p 0.05) for RASA. Members subject to a $500 annual deductible were more likely to experience high financial burden (adjusted odds ratio (AOR) = 1.677, p 0.001) and less likely to be adherent to diabetes medications (AOR = 0.484, p 0.001).The near-poor older adults enrolled in Wisconsin SeniorCare program had low financial burden and good medication adherence within the program.

The transition from prison to community is characterized by elevated morbidity and mortality, particularly owing to drug overdose. However, most formerly incarcerated adults with substance use disorders do not use any health care, including treatment for substance use disorders, during the initial months after incarceration.To evaluate whether a prerelease Medicaid enrollment assistance program is associated with increased health care use within 30 days after release from prison.This retrospective cohort study included 16 307 adults aged 19 to 64 years with a history of substance use who were released from state prison between April 1, 2014, and December 31, 2016. The Wisconsin Department of Corrections implemented prerelease Medicaid enrollment assistance in January 2015. Statistical analysis was performed from January 1 to August 31, 2021.A statewide Medicaid prerelease enrollment assistance program.The main outcome was Medicaid-reimbursed health care, associated with substance use disorders and for any cause, within 30 days of prison release, including outpatient, emergency department, and inpatient care. Mean outcomes were compared for those released before and after implementation of prerelease Medicaid enrollment assistance using an intention-to-treat analysis and person-level data from the Wisconsin Department of Corrections and Medicaid.The sample included 16 307 individuals with 18 265 eligible releases (men accounted for 16 320 of 18 265 total releases, and 6213 of 18 265 releases were among Black individuals; mean [SD] age at release, 35.5 [10.7] years). The likelihood of outpatient care use within 30 days of release increased after implementation of enrollment assistance relative to baseline by 7.7 percentage points for any visit (95% CI, 6.4-8.9 percentage points; P .001), by 0.7 percentage points for an opioid use disorder visit (95% CI, 0.4-1.0 percentage points; P .001), by 1.0 percentage point for any substance use disorder visit (95% CI, 0.5-1.6 percentage points; P .001), and by 0.4 percentage points for receipt of medication for opioid use disorder (95% CI, 0.2-0.6 percentage points; P .001). There was no significant change in use of the emergency department (0.7 percentage points [95% CI, -0.15 to 1.4 percentage points]). The probability of an inpatient stay increased by 0.4 percentage points (95% CI, 0.03-0.7 percentage points; P = .03).The results of this cohort study suggest that prerelease Medicaid enrollment assistance was associated with increased use of outpatient health care after incarceration and highlights the value of making this assistance universally available within correctional settings. More tailored interventions may be needed to increase the receipt of treatment for substance use disorders.

Context: States have experimented with the income eligibility threshold between Medicaid coverage and access to subsidized Marketplace plans in an effort to increase coverage for low-income adults while meeting other state priorities, particularly a balanced budget. In 2014, Wisconsin opted against adopting an ACA Medicaid expansion, instead setting the Medicaid eligibility threshold at 100% of the poverty level—a state-funded partial expansion. Childless adults gained new eligibility, while parents and caregivers with incomes between 101–200% of poverty lost existing eligibility. Methods: We used Wisconsin's all-payer claims database to assess health insurance gains, losses, and transitions among low-income adults affected by this partial expansion. Findings: We found that less than one third of adults who lost Medicaid eligibility definitely took up commercial coverage, and many returned to Medicaid. Among those newly eligible for Medicaid, there was little evidence of crowd-out. Both groups experienced limited continuity of coverage. Overall, new Medicaid enrollment of childless adults was offset by coverage losses among parents and caregivers, rendering Wisconsin's overall coverage gains similar to nonexpansion states. Conclusions: Wisconsin's experience demonstrates the difficulty in relying on the Marketplace to cover the near poor and suggests that full Medicaid expansion more effectively increases coverage.

Disruptions in care during the COVID-19 pandemic may have decreased access to care for patients with opioid use disorder.To examine trends in opioid use disorder treatment including buprenorphine possession, urine drug testing, and opioid treatment program services during the COVID-19 public health emergency.This cohort study included 6453 parent and childless adult Medicaid beneficiaries, aged 18 to 64 years, with opioid use disorder and continuous enrollment from December 1, 2018, to September 30, 2020, in Wisconsin. Logistic regression compared differences in study outcomes before, early, and later in the COVID-19 public health emergency. Analyses were conducted from January 2021 to October 2021.Early (March 16, 2020, to May 15, 2020) and later (May 16, 2020, to September 30, 2020) in the public health emergency.Person-week outcomes included possession of buprenorphine, completion of outpatient urine drug testing, and receipt of opioid treatment program services.The final cohort of 6453 participants included 3986 (61.8%) childless adults; 5741 (89%) were younger than 50 years, 3435 (53.2%) were women, 5036 (78.0%) White, and 22.0% were racial and ethnic minority groups (American Indian, 269 [4.2%]; Asian, 26 [0.4%]; Black, 458 [7.1%]; Hispanic, 292 [4.5%]; Pacific Islander, 1 [.02%]; Multiracial, 238 [3.7%]). Overall, 2858 (44.3%), 5074 (78.6%), and 2928 (45.4%) received buprenorphine, urine drug testing, or opioid treatment program services during the study period, respectively. Probability of buprenorphine possession did not change in the early or later part of the public health emergency. Probability of urine drug testing initially decreased (marginal effect [ME], -0.04; 95% CI, -0.04 to -0.03;In a sample of continuously enrolled adult Medicaid beneficiaries, the COVID-19 public health emergency was not associated with decreased probability of buprenorphine possession, but was associated with decreased probability of urine drug testing and opioid treatment program services. These findings suggest patients in office-based settings retained access to buprenorphine despite decreased on-site services like urine drug tests, whereas patients at opioid treatment programs experienced greater disruption in care. Given the importance of medications for opioid use disorder in preventing overdose, policy makers should consider permanent policy changes based on lessons learned from the public health emergency to enable ongoing enhanced access to these medications.

After the federal public health emergency was declared in March 2020, states could qualify for increased federal Medicaid funding if they agreed to maintenance of eligibility (MOE) provisions, including a continuous coverage provision. The implications of MOE provisions for total Medicaid enrollment are unknown.To examine observed increases in Medicaid enrollment and identify the underlying roots of that growth during the first 7 months of the COVID-19 public health emergency in Wisconsin.This population-based cohort study compared changes in Wisconsin Medicaid enrollment from March through September 2020 with predicted changes based on previous enrollment patterns (January 2015-September 2019) and early pandemic employment shocks. The participants included enrollees in full-benefit Medicaid programs for nonelderly, nondisabled beneficiaries in Wisconsin from March through September 2020. Individuals were followed up monthly as they enrolled in, continued in, and disenrolled from Medicaid. Participants were considered to be newly enrolled if they enrolled in the program after being not enrolled for at least 1 month, and they were considered disenrolled if they left and were not reenrolled within the next month.Continuous coverage provision beginning in March 2020; economic disruption from pandemic between first and second quarters of 2020.Actual vs predicted Medicaid enrollment, new enrollment, disenrollment, and reenrollment. Three models were created (Medicaid enrollment with no pandemic, Medicaid enrollment with pandemic economic circumstances, and longer Medicaid enrollment with a pandemic-induced recession), and a 95% prediction interval was used to express uncertainty in enrollment predictions.The study estimated ongoing Medicaid enrollment in March 2020 for 792 777 enrollees (mean [SD] age, 20.6 [16.5] years; 431 054 [54.4%] women; 213 904 [27.0%] experiencing an employment shock) and compared that estimate with actual enrollment totals. Compared with a model of enrollment based on past data and incorporating the role of recent employment shocks, most ongoing excess enrollment was associated with MOE provisions rather than enrollment of newly eligible beneficiaries owing to employment shocks. After 7 months, overall enrollment had increased to 894 619, 11.1% higher than predicted (predicted enrollment 805 130; 95% prediction interval 767 991-843 086). Decomposing higher-than-predicted retention, most enrollment was among beneficiaries who, before the pandemic, likely would have disenrolled within 6 months, although a substantial fraction (30.4%) was from reduced short-term disenrollment.In this cohort study, observed increases in Medicaid enrollment were largely associated with MOE rather than new enrollment after employment shocks. Expiration of MOE may leave many beneficiaries without insurance coverage.

The onset of a disability or major health shock can affect the labor supply of not only those experiencing the event but also their family members. Potential caregivers face a tradeoff between time spent earning income for the family and providing care for their spouse, which could be affected by the availability of paid leave. We examine caregiving and labor supply decisions after a spouse’s disability or health shock and the role of paid leave laws implemented in California and New Jersey in the response using data from the Survey of Income and Program Participation (SIPP). We show that labor force participation of potential caregivers decreased after spousal work-limiting disability or chronic health condition and, to a lesser extent, work-limiting illness. We find that paid leave reduces the likelihood that potential caregivers decrease their work hours to provide caregiving to their spouse after a work-limiting disability or chronic health condition, but limited evidence of effects on other employment outcomes. Our findings demonstrate that spousal disability and health shocks have long-run effects on household labor supply and therefore could be mediated by paid leave; we conclude by discussing possible reasons for finding limited impact in this context. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Disability onset and major health shocks can affect the labor supply of those experiencing the event and their family members, who face a tradeoff between time spent earning income and providing care. This decision could be affected by the availability of paid family leave. We examine the role of paid leave mandates in caregiving and labor supply decisions after a spouse's disability or health shock. Using data from the Survey of Income and Program Participation, we show that paid leave mandates reduce the likelihood that potential caregivers report decreasing their paid work hours to provide caregiving after a spouse's health shock. However, if caregivers are unlikely to have access to job protection, paid leave mandates also increase the likelihood of leaving the labor market to provide caregiving and working fewer weeks. There is limited evidence of an effect of paid leave on other employment outcomes. Our findings demonstrate that paid leave has some influence on household labor supply decisions after spousal health shocks, but its role should be considered together with the availability of job protection.

RESEARCH OBJECTIVE: Evidence suggests that expanding eligibility for in‐kind public welfare programs may improve health, economic, and crime‐related outcomes for former prisoners. The Affordable Care Act increased benefits for this population by authorizing expanded Medicaid eligibility. Coverage may increase access to treatment for conditions that impede employment and increase risk of recidivism (a health channel). By improving financial security, coverage may alter the incentives to commit crime (a financial channel). We estimate the effects of Medicaid coverage on employment and recidivism for adults released from state prison. STUDY DESIGN: We study sequential natural experiments in Wisconsin that expanded Medicaid availability to released prisoners, expansion of Medicaid to childless adults with income below 100% FPL, and the introduction of prison‐based Medicaid enrollment assistance. Using person‐level longitudinal data that links Corrections, Medicaid, and Unemployment Insurance data, we implemented two‐stage least squares instrumental variables (IV) analysis. Three variables comprised the key instruments corresponding to three policy periods: eligibility expansion; a 3‐month enrollment assistance implementation period; and the fully operational enrollment assistance program. The key assumption is that timing of release is unrelated to the outcomes, conditional on the other variables in the model. Outcomes included reincarceration within 6‐ and 12‐months, and post‐release quarterly employment status and earnings. POPULATION STUDIED: The population includes adults ages 19–64 incarcerated by the state who were released to the community between January 2013 – June 2017 (N = 32,846 individuals). The primary sample includes all releases; the secondary sample includes the first release per person. PRINCIPAL FINDINGS: The instruments were highly correlated with Medicaid enrollment in the month of release (F‐statistic of 7104). Relative to the baseline period, the percentage point (pp) increase in the likelihood of enrollment associated with expanded eligibility, the program implementation period, and the fully operational enrollment assistance program were 30.1, 47.6 and 61.1 respectively (p

Introduction: Many older adults face difficulty in affording their prescription drugs, despite having coverage available through Medicare Part D. SeniorCare is Wisconsin’s pharmaceutical assistance program that provides comprehensive drug coverage for low-income older adults who are not eligible for full Medicaid benefits. Methods: We analyzed SeniorCare enrollment and pharmacy claims data from 2014 to 2018. Results: Total drug expenditures increased by 19.3%, with the proportion of expenditures paid by SeniorCare and members decreasing while the proportion paid by other payers increased. Specialty drugs accounted for a substantial and growing proportion of total expenditures (20.4% in 2018) despite accounting for

Objective To estimate the association between the implementation of parity in coverage for mental health and substance use disorder (MHSUD) services within the Medicaid program and MHSUD service use. Data sources/study setting Wisconsin Medicaid enrollment and claims data from 2013 to 2015. In April 2014, Wisconsin Medicaid transitioned childless adult beneficiaries from coverage with limited MHSUD services to parity-consistent coverage. Preparity, they only had Medicaid coverage for MHSUD visits to psychiatrists and the emergency department, while parent beneficiaries had parity-consistent coverage. Study design The study uses a difference-in-differences design to compare outcome changes for childless adult and parent beneficiaries. Data collection/extraction methods We identified 76, 569 childless adult and parent beneficiaries aged 18-64 who were continuously enrolled for the 2-year study period. Principal findings Introducing parity-consistent coverage within Medicaid was associated with increased utilization of Medicaid-reimbursed MHSUD services: outpatient, prescription medication, ED, and inpatient. Increased MHSUD outpatient visits were driven by increased visits to nonpsychiatrists. Conclusions Parity's effects on MHSUD service use have been studied in the context of private insurance, but its impact among Medicaid beneficiaries has not. Our findings suggest that parity implementation in Medicaid could increase access to effective MHSUD services in a high-need population.

This paper discusses causal inference techniques for social scientists through the lens of applied microeconomics. We frame causal inference using the standard of the ideal experiment, emphasizing problems of omitted variable bias and reverse causality. We explore how laboratory and field experiments can succeed and fail to meet this ideal in practice. We then outline how different methods and the statistical assumptions behind them can lead to causal inference in non-experimental contexts. We explain when problems with omitted variables bias can and cannot be addressed using observed controls. We consider tools for studying natural experiments, including difference-in-differences, instrumental variables, and regression discontinuity techniques. Finally, we discuss additional concerns that may arise such as weighting, clustering, multiple inference, and external validity. We include Stata code for implementing each of these methods as well as a series of checklists for researchers detailing important robustness and design checks. Throughout, we emphasize the importance of understanding the context of a study and implementing analyses in a way that acknowledges strengths and limitations.

This paper discusses causal inference techniques for social scientists through the lens of applied microeconomics. We frame causal inference using the standard of the ideal experiment, emphasizing problems of omitted variable bias and reverse causality. We explore how laboratory and field experiments can succeed and fail to meet this ideal in practice. We then outline how different methods and the statistical assumptions behind them can lead to causal inference in non-experimental contexts. We explain when problems with omitted variables bias can and cannot be addressed using observed controls. We consider tools for studying natural experiments, including difference-in-differences, instrumental variables, and regression discontinuity techniques. Finally, we discuss additional concerns that may arise such as weighting, clustering, multiple inference, and external validity. We include Stata code for implementing each of these methods as well as a series of checklists for researchers detailing important robustness and design checks. Throughout, we emphasize the importance of understanding the context of a study and implementing analyses in a way that acknowledges strengths and limitations.

Objective This study measures the change in health care use after enrollment into a new public insurance program for low-income childless adults. Data Sources/Study Setting The data sources include claims from a large integrated health system in rural Wisconsin and Medicaid enrollment files, January 2007–September 2012. Study Design We employ a regression discontinuity design to measure the causal effect of public insurance enrollment on counts of outpatient, emergency department, and inpatient events for 2 years following enrollment for a sample of previously uninsured low-income adults in rural Wisconsin. Principal Findings Public insurance enrollment led to substantial increases in outpatient visits including preventive visits, but not mental health visits. Public insurance enrollment also led to increases in inpatient stays, but the study is inconclusive on whether it led to an increase in ED visits. Conclusions Public insurance expansions to childless adults have the potential to impact the use of health care. The large increase in Medicaid coverage and reduction in rates of uninsurance anticipated to result from the Affordable Care Act should increase the use of inpatient and outpatient services, but they will have an uncertain impact on the use of ED among rural populations.

This survey study compares the demographic characteristics, access to care, and health status of recent Medicaid enrollees in Wisconsin with and without experience of Medicaid lockouts.

Wisconsin's BadgerCare Plus (BC+) program was designed to ensure access to health insurance coverage to virtually all Wisconsin children and to bolster coverage for parents and other caretaker adults. The program, launched in February 2008, expanded upon BadgerCare (Wisconsin's Children's Health Insurance Program) and Medicaid. BC+ combines its predecessor programs into one streamlined umbrella program and its introduction was characterized by the following reforms: an eligibility expansion augmented with an autoenrollment initiative; simplification of eligibility rules and enrollment and verification processes; and an aggressive marketing and outreach campaign. This paper provides estimates of the impact of the reform effort on program enrollment and the continuity of coverage. We find that the introduction of BC+ was associated with increases in enrollment and decreases in exit rates, suggesting that the Wisconsin experience may be instructive for other states interested in achieving similar aims.

We study the welfare consequences of a switch from risk-rated premiums to community-rated premiums using a two stage model of health insurance demand and detailed claims data from Chilean health insurers. In this context, because the consumer's available choice set depends directly on income, effectively creating a minimum required plan, classical predictions of adverse selection models (in particular, unraveling of the market) may not obtain. Our results suggest that if insurers are not allowed to scale premiums based on age and sex (in accordance with a recent court ruling), a significant fraction of young, healthy consumers will opt out of the private insurance market, raising the average cost of insuring those who remain. Chilean consumers in aggregate are predicted to be worse off under community-rated premiums, with a total decline in consumer surplus of 1.3%. This result remains even when we do not allow consumers to choose their plans based on private information. While consumers may perceive uniform prices as more equitable, they are less efficient in this market.

The Patient Protection and Affordable Care Act (ACA) relies heavily on the expansion of Medicaid eligibility to cover uninsured populations. In February 2008, Wisconsin expanded and reformed its Medicaid/CHIP program and, as part of program implementation, automatically enrolled a set of newly eligible parents and children. This process of “auto-enrollment” targeted newly eligible parents and older children whose children/siblings were already enrolled in the state's Medicaid/CHIP program. Auto-enrollment brought over 44,000 individuals into the program, representing more than 60% of all enrollees in the first month of the reformed program. Individuals who were auto-enrolled were modestly more likely to leave the program relative to other individuals who enrolled in February 2008, unless their incomes were high enough to be required to pay premiums; these auto-enrollees were much more likely to exit relative to other enrollees subject to premium payments. The higher exit rates exhibited by non-premium paying auto-enrollees were likely due to the fact that over 40% of auto-enrollees were covered by a private insurance policy in the month of their enrollment, compared to approximately 30% for regular enrollees. A national simulation of an auto-enrollment process similar to Wisconsin's, including the expansion of adult Medicaid eligibility to 133% of the federal poverty level under the ACA, suggests that 2.5 million of the 5.6 million newly eligible parents could be auto-enrolled, and approximately 25% of this population would be privately insured. These results suggest that auto-enrollment may be appropriate for other states, especially in their efforts to enroll eligible populations who are not subject to premium requirements.

This study provides plausibly causal estimates of the effect of public insurance coverage on the employment of nonelderly, nondisabled adults without dependent children (“childless adults”). We use regression discontinuity and propensity score matching difference-in-differences methods to take advantage of the sudden imposition of an enrollment cap, comparing the labor supply of enrollees to eligible applicants on a waitlist. We find that enrollment into public insurance leads to sizable and statistically meaningful reductions in employment up to at least nine quarters later, with an estimated size of 2-10 percentage points, depending on the model used.

This paper estimates the effect that premiums in Medicaid have on the length of enrollment of program beneficiaries. Whether and how low income-families will participate in the exchanges and in states’ Medicaid programs depends crucially on the structure and amounts of the premiums they will face. I take advantage of discontinuities in the structure of Wisconsin's Medicaid program to identify the effects of premiums on enrollment for low-income families. I use a 3-year administrative panel of enrollment data to estimate these effects. I find an increase in the premium from 0 to 10 dollars per month results in 1.4 fewer months enrolled and reduces the probability of remaining enrolled for a full year by 12 percentage points, but other discrete changes in premium amounts do not affect enrollment or have a much smaller effect. I find no evidence of program enrollees intentionally decreasing labor supply in order to avoid the premiums.

As states consider expanding Medicaid to low-income childless adults under the Affordable Care Act, their decisions will depend, in part, on how such coverage may affect the use of medical care. In 2009 Wisconsin created a new public insurance program for low-income uninsured childless adults. We analyzed administrative claims data spanning 2008 and 2009 using a case-crossover study design on a population of 9,619 Wisconsin residents with very low incomes who were automatically enrolled in this program in January 2009. In the twelve months following enrollment in public insurance, outpatient visits for the study population increased 29 percent, and emergency department visits increased 46 percent. Inpatient hospitalizations declined 59 percent, and preventable hospitalizations fell 48 percent. These results demonstrate that public insurance coverage expansions to childless adults have the potential to improve health and reduce costs by increasing access to outpatient care and reducing hospitalizations.

We use administrative data from Wisconsin to estimate the percent of individuals newly enrolled in public health coverage that had access to private, employer-sponsored health insurance at the time of their enrollment and the percent that was uninsured. We also estimate the percent of new enrollees that dropped private coverage in the six months prior to or following enrollment. We estimate that, among all new enrollees and after expansion of eligibility for public coverage, approximately 21% had access to private health insurance at the time of enrollment and that only 10% dropped this coverage. We also estimate strict bounds for these percentages and find that the percentage of new enrollees with private insurance coverage at the time of enrollment lies between 16% and 29% and the percentage that dropped private coverage in favor of public insurance lies between 4% and 18%. These estimates of crowd-out are relatively low compared with estimates from the literature based on Medicaid and Children’s Health Insurance Program expansions, although based both on different data and on a different method. In order to facilitate comparison of our method with previous studies, we develop a conceptual method for understanding the various types of crowd-out measures.

We use a combination of administrative and survey data to estimate the fraction of individuals newly enrolled in public health coverage (Wisconsin's combined Medicaid and CHIP program) that had access to private, employer-sponsored health insurance at the time of their enrollment and the fraction that dropped this coverage. We estimate that after expansion of eligibility for public coverage, approximately 20% of new enrollees had access to private health insurance at the time of enrollment and that only 8% dropped this coverage (with the remaining 12% having both private and public coverage). We also identify an "upper bound" estimate, which suggests that the percentage of new enrollees with private insurance coverage at the time of enrollment is, at most, 27%. These estimates of crowd-out are relatively low compared with estimates from the literature based on Medicaid and CHIP expansions, although based both on different data and on a different method.

We use administrative data from Wisconsin to determine the fraction of new Medicaid enrollees who have private health insurance at the time of enrollment in the program. Through the linkage of several administrative data sources not previously used for research, we are able to observe coverage status directly for a large fraction of enrollees and indirectly for the remainder. We provide strict bounds for the percentages in each status and find that the percentage of new enrollees with private insurance coverage at the time of enrollment lies between 16 percent and 29 percent, and the percentage that dropped private coverage in favor of public insurance lies between 4 percent and 18 percent. Our point estimates indicate that, among all new enrollees, 21 percent had private health insurance at the time of enrollment and that 10 percent dropped this coverage. Our results show substantially lower rates than previous studies of crowd-out following public health insurance expansions and significant rates of dual coverage, whereby new enrollees into public insurance retain their previously held private insurance coverage.