Sen. Lindsey O. Graham (R-S.C.). (Andrew Harrer/Bloomberg)

Timothy Jost is an emeritus professor at Washington and Lee University law school.

The Senate is on the verge of launching a dangerous experiment. Having failed for months to repeal and replace the Affordable Care Act, Republicans hope to dump onto the states the problem of providing low- and moderate-income Americans with access to health care. So long as they are determined to push forward without meaningful Democratic input, they have until Sept. 30 to act under Senate rules. The Graham-Cassidy bill — their last chance to meet this deadline — would simply hand the states block grants and abandon to them millions of Americans whom the ACA now helps.

In doing so, however, Graham-Cassidy would radically redistribute health-care funding. It would effectively strangle the efforts of states — predominantly Democratic — that have been most successful in extending coverage under the ACA’s Medicaid expansion, tax credits and cost-sharing reductions. Conversely, it would lavish funding on states — predominantly Republican — that have resisted the coverage expansion opportunities the ACA offered. It would also allow states to waive the basic health insurance protections the ACA has given individuals with preexisting conditions. Finally, like previous Republican bills, it would cap federal Medicaid funding growth at levels that would force states to cut back on traditional Medicaid coverage, further increasing the numbers of the uninsured.

If passed, Graham-Cassidy would proceed in three stages. The current ACA premium tax credits and Medicaid expansion would remain in place for 2018 and 2019. But the bill would immediately — indeed retroactively — end the individual- and employer-mandate penalties, while offering nothing in return to encourage healthy enrollees to maintain coverage. The bill would also do nothing to compensate insurers for the reductions they must make in deductibles and other cost-sharing for low-income enrollees, an issue that is currently driving up premiums. The bill includes some funding that could be used to help stabilize insurance markets, but does not specify how the money would be used, and in any event, it provides too little money. Under this law, it is quite possible that the individual insurance markets could collapse in some states before 2020.

In 2020, the second stage of Graham-Cassidy would go into effect. Between 2020 and 2026, funding the states currently receive would be reallocated according to a complex formula that would, subject to several adjustments, base each state’s share on its portion of the U.S. population with incomes between 50 and 138 percent of the federal poverty level. While the sponsors of the bill assert that this is fair, it would create numerous problems.

To begin, it would dramatically cut funding for states that have successfully extended coverage under the ACA. California would effectively receive $28 billion less in 2026 than under currently projected funding, New York $19 billion less. On the other hand, the bill would just as dramatically expand funding for states that have largely ignored their uninsured: Funding for Texas would increase $8 billion over currently projected levels.

While additional funds could allow ACA-resistant states to do more to cover their uninsured, states would first have to establish programs to distribute the money. This would likely require state legislation, administrative rules and possibly the establishment of new agencies. And state action would have to await for federal agencies to first adopt their own rules implementing the bill. It is hard to see how this could all take place between now and 2020.

Further, Graham-Cassidy would allow states to waive the ACA requirement that insurers must cover essential health benefits, such as prescription drugs or mental health and substance-use disorder services, as well as the ACA’s prohibition against insurers charging higher premiums to people with preexisting conditions. States would have to describe how they intend “to maintain access to adequate and affordable health insurance coverage for individuals with pre-existing conditions,” but it is unclear what “adequate and affordable” means and difficult to understand how coverage could meet this standard if services to treat preexisting conditions are not available or if individuals with preexisting conditions can be charged higher premiums. Moreover, the bill contains no mechanism for revoking funding to a state if it fails in fact to meet its commitments.

Graham-Cassidy allows states to spend up to 15 percent of their grants on covering traditional Medicaid enrollees through private insurance. States could simply replace their traditional Medicaid expenditures with Graham-Cassidy funding or use the funds to make up for Medicaid shortfalls as other provisions in the bill limit federal funding of traditional Medicaid. This would reduce the money available to cover the low- and moderate- income people that the ACA now covers.

Finally, the entire block-grant program depends on private insurer participation. But insurers would face 50 different state regulatory and financing systems, each of which could change from year to year. Many private insurers may decide this is just too risky a market to play in.

The third stage of Graham-Cassidy would arrive in 2027, as the bill’s appropriations and funding formulas expire. Congress would have to pass legislation essentially establishing a new program at that point. And it would have to pay for it in some way. Any states that expanded their programs in response to Graham-Cassidy would have to embrace the possibility of massive funding shortfalls at that time. A very risky experiment indeed.

Americans have decisively rejected Republican schemes to repeal the Affordable Care Act all summer. They should tell the Senate to reject Graham-Cassidy as well.