Bernie Sanders’s high-profile campaign for Medicare for All has pushed the Democratic Party’s conversation about health care reform to the left and opened up space for center-left Democrats to embrace proposals that might have seemed radical just a few years ago.

But while Sanders has made the proposal the centerpiece of his run for president, in 2017, he rejected the idea that support for his approach should be a litmus test for Democrats.

Many others on the left have done just that. But it’s important to be realistic about his bill’s chances of becoming law if the party wins big in 2020. Even if Democrats end up with a significant majority in the House and more than 50 seats in the Senate—and kill the filibuster—they won’t end up with progressive majorities in both chambers who will vote for Medicare for All.

A big reason for that is that public opinion surveys consistently find that mandating that everyone shift into Medicare, as Sanders would do, is significantly less popular than proposals for allowing individuals and businesses to opt in to the program voluntarily.

Both approaches, if designed well, could achieve the same broad goals—delivering comprehensive, universal coverage at lower costs to households, and better controlling system-wide costs. It’s hard to imagine centrist and center-left Dems rallying around Sanders’s approach when voters prefer a voluntary scheme that would come with a smaller price tag for the federal government and be easier to sell politically.

But an optional buy-in could be transformational only if it’s designed properly. The devil is in the details.

The Democratic presidential candidates have floated a number of different proposals for a “public option” approach to reform, but they vary widely in their ambition. So it’s important for progressives to understand the details—and, if Democrats end up with unified control in 2020, to be prepared to walk away from the negotiating process if key conditions aren’t met.

Universality should be first and foremost among these conditions—including immediately covering the estimated 4.9 million low-income people who were screwed as a result of red states’ refusing to expand their Medicaid programs. Reforms must also include assuring that vital benefits such as mental health care and prescription drugs are covered, and it must limit the costs of premiums, co-pays, and other out-of-pocket expenses so people don’t go bankrupt paying for necessary care.

Each of the five leading candidates, with the exception of Mayor Pete Buttigieg, has a major health care reform plan as part of their platform. In press interviews and a brief section on his website, Buttigieg has proposed a straightforward public option that would allow everyone under age 65 to buy into a Medicare-type plan at cost. (He separately proposed new incentives to expand access to health care in rural America.) This is the most modest iteration of a “public option,” and would leave Americans uninsured who can’t afford coverage or choose not to buy in. Everyone else in the top tier, including former vice president Joe Biden, has touted more aggressive reforms.

Biden’s plan is surprisingly ambitious for a candidate running as a moderate alternative to Trump. It would add a new public insurance option offered through the Affordable Care Act (ACA) exchanges, and would automatically enroll people in non-expansion states who make less than 138 percent of the federal poverty line. They would not be required to pay premiums. Everyone else—individuals and businesses—would have the option of buying in to the new public alternative, and Biden would contain their costs by significantly expanding the ACA’s subsidies and making them available to households whose income levels makes them ineligible for assistance now.

According to his campaign’s own estimate, his proposal would reduce the number of uninsured to 3 percent but ultimately fail the test of universality. And while Biden recently reversed his support for the Hyde Amendment—which bars federal funding for most abortions—and would cover all reproductive health care in his public option, his plan would not mandate coverage for mental health, dentistry, or vision care. (He vows to “redouble” his past “efforts to ensure enforcement of mental health parity laws and expand funding for mental health services.”)

Senator Elizabeth Warren has endorsed Sanders’s plan. It passes these tests with flying colors, offering everyone public insurance with more expansive benefits than is typical for countries with single-payer systems. There would be no premiums or co-pays.

Senator Kamala Harris proposes getting to a Medicare for All system that is closer to the existing Medicare program in that private insurers would still be able to sell highly regulated Medicare-compliant plans. Like Sanders’s plan, it would improve Medicare’s coverage to include dental, vision, prescription drugs, and hearing aids. Her proposal envisions a 10-year transition period to get there. (Sanders’s plan would phase in over four years.)

Harris’s transitional plan appears to be modeled on the Medicare for America Act, a bill introduced in the House by Representative Rosa DeLauro (D-CT) and over a dozen co-sponsors. It’s a plan that’s worth considering in detail, because it is also a comprehensive, transformative proposal that passes these litmus tests and might attract support from the Democratic Party’s different factions.

It shares three vitally important features of Medicare for All, all of which distinguish it from more modest “public option” schemes. First, like Sanders’s bill, Medicare for America would offer automatic enrollment. Every newborn would be enrolled in the program at birth, as would everyone who is currently uninsured. If someone without coverage requires care, they could simply go to a doctor, clinic, or hospital, get covered and receive treatment. Their premiums, calculated on a sliding scale, would be added to their federal taxes. Anyone whose income is under 200 percent of the federal poverty line would not have to pay for their coverage.

This is important for achieving a universal system. There’s a common belief that in countries with single-payer systems, the government simply covers everyone out of general tax revenues, but that’s a misconception. As Yale political scientist Jacob Hacker told me in an interview a couple of years back, “in other countries, you’re basically guaranteed coverage and then they figure out how to pay for it. Some of that money may come from you, some will come from your employer and some of it will come from general funds. We don’t have that approach. People who don’t have coverage from their employers have to figure out how to sign up—either for Medicaid, or through the exchanges.” Medicare for America and Medicare for All would both bring that cover-first approach to our system.

Second, Medicare for America would roll Medicaid and the Children’s Health Insurance Program, both of which are administered by the states with federal oversight, into the Medicare program. This provision, also in the Sanders bill and Harris’s transition plan, is vitally important not only because 14 red states continue to refuse the Affordable Care Act’s Medicaid expansion, but also because those states generally tend to have narrower eligibility requirements and stingier benefits than their blue counterparts.

Harris’s and Sanders’s proposals and Medicare for America would all be financed in large part with new taxes on the wealthiest Americans. They vary in the specifics, but they share the philosophy of using the system to make those at the top pay their fair share.

Medicare for America and Harris’s transitional plan both control costs to individuals by increasing premium subsidies and placing caps on out-of-pocket expenses. Harris’s campaign has said she’d cap out-of-pocket expenses at $200 per year, but offers no other details. Under DeLauro’s proposal, those making less than twice the federal poverty line, or who don’t have to pay federal income taxes, wouldn’t have to pay premiums. People who make more than that would have theirs subsidized on a sliding scale up to six times the poverty line, or around $150,000 for a family of four.

The three proposals differ in that Sanders’s Medicare for All proposal would eliminate almost all private insurance, Harris’s plan would allow private insurers to sell policies through the government-run program, and DeLauro’s bill would allow businesses to continue to offer private insurance outside of Medicare as long as it is “gold-level” coverage.

Over the long run, advocates of this kind of optional approach believe that most people would end up in Medicare because it is more efficient than private insurers and would contain costs by setting prices for a very large pool of insured. That would make it more attractive than what private insurers could offer, and since newborns would be covered and people would be automatically enrolled in Medicare when they lose other coverage, over time the system would end up looking like Medicare for All. (If Sanders’s proposal were implemented, some people—veterans, those who receive care from the Indian Health Service and those who require long-term care under Medicaid—would also remain covered by a payer other than Medicare.)

Despite the usual hand-wringing of pundits and moderate Democrats about the party moving too far to the left on health care, Democrats are in a good place heading into 2020. Republicans have thoroughly discredited themselves on this issue with repeated, highly unpopular attempts to strip coverage from people and kill protections for those with preexisting conditions. As for the Dems, after pre-compromising on the Affordable Care Act, even the moderate candidates are pushing for some relatively ambitious reforms.

The left should continue to make demands based on our values, but in the end we need to make sure that when legislators meet in the middle it ends up more like Medicare for America or Kamala Harris’s proposal than what Pete Buttigieg and Joe Biden are offering.