In 2016, Medicaid and the Children’s Health Insurance Program (CHIP) spent $582 billion in federal and state tax dollars. This represented nearly one in five dollars spent on healthcare in the United States. Taxpayers provide massive monetary inputs for these medical welfare programs, but what about the outcomes?

Although CHIP and Medicaid were implemented decades ago (1998 and 1965 respectively), researchers have conducted relatively few studies testing their effects on participants’ health outcomes. Furthermore, the results from these few studies are mixed.

CHIP, formerly called SCHIP, is a health insurance program for children whose families have too much income to qualify for Medicaid. In 2012, Dr. Embry Howell, a health policy fellow at the left-leaning Urban Institute think tank, compiled the six national studies that tracked health outcomes for children on CHIP since its implementation in 1998. Dr. Howell’s analysis left her with disparate results. In her report, she concluded, “…the limited evidence of the impact of the Medicaid/CHIP expansions on child health is mixed, with as many studies showing no change as showing positive improvements in health.”

The results were similar in nature for researchers studying the health outcomes of Medicaid beneficiaries.

In 2008, Oregon expanded its Medicaid program to 10,000 low-income individuals using a lottery system. Economists and health care experts from Harvard and MIT recognized this was the perfect setting to conduct the first large-scale randomized control trial (accepted as the “gold standard” in medical research) on the effects of Medicaid enrollment on health outcomes.

Using an array of metrics, researchers compiled health outcomes for two years for Medicaid lottery winners and compared them to those who were not selected by the Medicaid lottery. After two years, researchers found “no significant improvements in measured physical health outcomes in the first 2 years….” This result sparked a backlash among advocates for Medicaid expansion or other programs to increase government involvement in healthcare.

Following the Oregon study, research from the University of Virginia, Johns Hopkins Hospital, and the University of Pittsburgh Cancer Institute showed Medicaid patients who had lung cancer, transplant surgeries, or other major surgical operations actually fared far worse than those who were privately insured or had no insurance at all.

It should be noted that although these three observational studies found distinct correlations between Medicaid patients and poor health outcomes, the authors of the studies were hesitant to assert Medicaid causes poor outcomes. They disclosed limitations in their studies such as possible selection bias and not fully controlling for certain variables, including socioeconomic status between the observed groups.

Medicaid and CHIP cost more than half a trillion dollars a year. The structure of these programs—taking money out of the states as federal taxes, then offering it back to states only if they abide by federal program rules—forces states to participate and prevents states from freely experimenting to try to get better results.

With taxpayers spending more than half a trillion dollars every year, the lack of research, let alone “gold standard” studies, on whether the programs work is astonishing. Policymakers should test outcomes in order to know what effects, if any, these programs have on those they are designed to help. Congress and state legislatures need to conduct oversight, which includes measuring the performance of welfare programs. Evaluating the effectiveness of these programs is impossible to do in the dark.