Does Medicaid Managed Care Improve Costs and Quality Through Private Competition?

Shepard and Wallace (2026) ask how states should design Medicaid managed care to balance costs, quality, and competition. They review evidence from prior causal studies, Medicaid administrative data, enrollment and spending trends, and cross-state variation in procurement and plan design. They find that managed care can reduce health care use and spending but has not reliably lowered government costs or improved quality. One randomized Louisiana study found 7 percent lower total health care spending and 25 percent lower pharmacy spending under managed care. They conclude that outcomes depend heavily on procurement rules, insurer incentives, market structure, and enrollee choice mechanisms.

Why This Article Was Selected for The Policy Scientist

This article addresses a major question in public policy: whether governments can use private organizations to deliver essential services while controlling costs and maintaining quality. That issue extends well beyond Medicaid to education, corrections, transportation, and other contracted public programs. The topic is especially timely because Medicaid now relies heavily on managed care, while states continue to face substantial fiscal and administrative pressures. Shepard and Wallace, both established contributors to health-policy research, provide a useful synthesis that connects classic work on contracting and incomplete contracts to modern Medicaid institutions. Published in the Journal of Economic Perspectives, one of economics’ leading journals for authoritative syntheses, the article benefits from broad administrative and cross-state evidence, although much of the underlying literature varies in data quality and external validity. Its strongest evidence comes from causal studies, including random assignment designs; future work would benefit from additional randomized or quasi-experimental evaluations of procurement and market-design reforms.

Full Citation and Link to Article

Shepard, M., & Wallace, J. (2026). Understanding Medicaid managed care: The procured competition model. Journal of Economic Perspectives, 40(2), 171–194. https://www.aeaweb.org/articles?id=10.1257%2Fjep.20251475

Central Research Question

The article examines how states should design Medicaid managed care to balance government costs, health care spending, access, quality, and competition among private insurers. The authors argue that the conventional question—whether Medicaid should be publicly administered or contracted to private managed care organizations—is increasingly less relevant because most states have already committed to managed care. By 2024, approximately 85 percent of Medicaid beneficiaries were enrolled in some form of managed care. The more consequential question is therefore how states should structure the system once private insurers participate.

The authors characterize Medicaid managed care as a system of “procured competition.” It combines elements of traditional government procurement with regulated market competition. States select a limited number of insurers through procurement, specify many of the terms under which those insurers operate, establish payment arrangements, and monitor performance. Beneficiaries then choose among the selected insurers, although many are automatically assigned to plans. The central analytical problem is whether this hybrid system creates appropriate incentives for insurers to reduce unnecessary spending while preserving access to necessary care and maintaining quality.

Previous Literature

The conceptual foundation begins with the literature on government contracting and incomplete contracts. Hart, Shleifer, and Vishny (1997) demonstrate that privatization can produce stronger incentives for cost reduction but can also lead contractors to reduce dimensions of quality that governments cannot easily observe or specify in contracts. Their framework is particularly relevant to health insurance because determining whether an insurer has provided adequate access and appropriate care is substantially more difficult than specifying the characteristics of many conventionally procured goods.

The article also draws on Laffont and Tirole’s (1993) work on incentives in procurement and regulation and Shleifer’s (1998) analysis of privatization. These studies emphasize the trade-off between high-powered private incentives to control costs and the possibility that private contractors will pursue savings in socially undesirable ways.

A substantial empirical literature evaluates whether Medicaid managed care actually reduces spending. Duggan (2004) examined Medicaid contracting in California, while Duggan and Hayford (2013) studied state and local managed-care mandates. Herring and Adams (2011) examined utilization under Medicaid HMOs. More recent causal studies provide stronger identification. Dranove, Ody, and Starc (2021) evaluated changes in the treatment of prescription drugs under managed care. Geruso, Layton, and Wallace (2023) exploited random assignment among Medicaid plans to estimate differences in plan spending. Macambira et al. (2025) used random assignment between managed care and fee-for-service Medicaid in Louisiana.

Another strand of research focuses on quality and access. Kuziemko, Meckel, and Rossin-Slater (2018) examined infant health outcomes following Medicaid managed-care expansion in Texas. Kreider et al. (2024) studied insurer network design and adverse selection, while Wallace (2023) examined how provider-network breadth affects spending and enrollee satisfaction. Collectively, this literature establishes that Medicaid managed care changes insurer incentives substantially, but it provides no uniform conclusion that privatization improves both costs and quality.

Data

This is primarily a synthetic and conceptual article rather than a single original causal evaluation. The authors assemble evidence from numerous published studies and combine it with descriptive administrative data on Medicaid enrollment, spending, insurer participation, and state program design.

Historical enrollment and expenditure figures draw on data from the Medicaid and CHIP Payment and Access Commission and the Kaiser Family Foundation. These data show Medicaid expanding from approximately 20 million beneficiaries in the early 1970s to more than 90 million during the COVID-era enrollment peak, before declining to about 77 million by late 2025. Inflation-adjusted federal and state Medicaid spending rose from roughly $100 billion in the 1970s to more than $900 billion annually.

The analysis of managed-care market structure also uses December 2022 T-MSIS Analytic Files for 36 states with sufficiently complete comprehensive managed-care data. These records permit comparisons of the number of managed care organizations available across counties and states. Procurement practices are supplemented with state contracting records, prior studies of Medicaid procurement, and examples from Oregon, California, Florida, Minnesota, and Nebraska.

Methods

The article does not estimate a single statistical model. Instead, the authors develop a conceptual framework and synthesize evidence from studies using different research designs. They distinguish between government spending—the amount states pay managed-care organizations—and actual health care spending on services used by beneficiaries. This distinction is essential because reduced medical utilization does not necessarily generate equivalent taxpayer savings.

Particular emphasis is placed on studies with credible causal identification. Macambira et al. exploit random assignment of Louisiana Medicaid beneficiaries between managed care and fee-for-service coverage. Geruso, Layton, and Wallace use random plan assignment to identify variation in insurer spending. Dranove, Ody, and Starc exploit state policy changes that moved prescription-drug benefits into managed care. Other cited studies use mandates, policy transitions, plan exits, and regulatory changes to identify effects.

The authors then organize Medicaid institutional design around three policy domains: procurement rules, market-design rules, and plan-choice rules. They use economic theories of procurement, incomplete contracts, adverse selection, risk adjustment, and consumer choice to interpret the empirical findings and identify unresolved research questions.

Findings/Size Effects

The clearest empirical finding is that Medicaid managed care sometimes reduces actual health care spending without correspondingly reducing government spending. Macambira et al. (2025), using random assignment in Louisiana, estimate that managed care reduced total health care spending by approximately 7 percent relative to fee-for-service Medicaid. Pharmacy spending declined by about 25 percent. Dranove, Ody, and Starc (2021) similarly estimate approximately a 21 percent reduction in prescription-drug spending following incorporation of pharmacy benefits into managed care.

However, studies examining government expenditures generally report effects near zero or increases in spending. Consequently, private plans may succeed in reducing medical expenditures without transferring the resulting savings to taxpayers. Administrative expenses, insurer profits, payment formulas, and subsequent adjustments to capitation rates may absorb part of the difference.

Insurers also differ substantially in their behavior. Geruso, Layton, and Wallace (2023) find approximately 25 percent variation in health care spending across Medicaid plans among beneficiaries assigned to different insurers. Importantly, beneficiaries tended to prefer the higher-spending plans, suggesting that spending differences may reflect meaningful differences in access or perceived quality rather than pure inefficiency.

Provider networks create another important trade-off. Research reviewed in the article indicates that narrower networks reduce spending but can restrict both necessary and unnecessary utilization and reduce beneficiary satisfaction. Kreider et al. (2024), for example, document a roughly 50 percent increase in cancer patients enrolling in a Medicaid plan after it added a prominent cancer hospital to its network. Because those patients were expensive and risk adjustment did not fully compensate the insurer, the plan subsequently removed the hospital.

Evidence regarding health outcomes and quality remains heterogeneous. Some studies identify worse birth outcomes, increased mortality, or lower satisfaction, while others find improved access, benefits, or no detectable harm. The authors therefore conclude that existing evidence does not establish a general causal relationship between Medicaid privatization and quality.

Conclusion

The article concludes that Medicaid managed care should not be understood simply as privatization. It is a hybrid institutional arrangement in which states first select private insurers and then regulate competition among them. Its performance therefore depends critically on procurement procedures, payment incentives, market rules, provider-network requirements, risk adjustment, and the mechanisms through which beneficiaries select or are assigned to plans.

Existing research provides reasonably strong evidence that managed care can reduce some categories of health care spending, but substantially weaker evidence that those reductions lower taxpayer expenditures or systematically improve quality. The principal unresolved issue is therefore institutional design rather than a categorical choice between public and private administration.

The large differences across states offer substantial opportunities for future causal research. Particularly valuable studies would examine changes in procurement scoring, competitive bidding, contract duration, automatic enrollment rules, risk-sharing provisions, and insurer participation. Randomized interventions where feasible, together with natural experiments and strong quasi-experimental designs, could determine which components of procured competition actually improve costs, access, and quality.