Democrats seem to agree: A job guarantee will be part of the 2020 election debate. What they don't yet agree on is what a "job guarantee" is.

Sen. Bernie Sanders (I-Vt.) is working on a full blown national bill to enshrine the right to a job. Sen. Cory Booker (D-N.J.) has a bill that would essentially run a test pilot of Sanders' program. The Congressional Full Employment Caucus has been pushing yet another national job guarantee bill, with some differences from Sanders' plan, for years. Finally, Rep. Ro Khanna (D-Calif.) just introduced his own bill, which would subsidize more private and public hiring alike.

The Democrats are a big diverse party, and they'll have to negotiate a lot of different preferences and approaches to ultimately unite behind such an ambitious idea. But as they come to an agreement around one shared definition of a job guarantee, the Democrats also need to understand the choices they make along the way — and their consequences.

Sanders sounds like he's working on a genuine guarantee: A government-run national system that, if you're unemployed and want work, will consult with your community to design a job for you that pays $15 an hour, plus health benefits. This would eradicate involuntary unemployment, and set a quality floor that no job — public or private — could fall below without simply losing work to the job guarantee program. That's especially important to people of color and other marginalized Americans, usually shoved to the fringes of employment: They're the last to benefit from business cycle upswings in private hiring, and the first to get tossed aside in downturns.

The drawback is that it's hugely ambitious. Sanders' job guarantee could expand the federal workforce by many millions. Plenty of left-wing economists, fully devoted to the goal of permanent full employment, nonetheless doubt the federal government could actually manage such a task.

Khanna's bill is a response to those concerns. Rather than bypassing the private sector with direct government hiring, Khanna would subsidize private and public employers alike to encourage them to hire more. The subsidies would support an unemployed or underpaid person getting a new job for 18 months, with some extensions available. And then there are some other limits and caveats.

Khanna's approach certainly wouldn't be bureaucratically simple. But it would work with employers who already exist rather than trying to build a whole new public option employer. (Though it's worth noting that different versions of the full-blown job guarantee also tackle this problem in interesting ways.)

The downside is that it's not really a job guarantee, a point Khanna himself freely admits. Rather, it tries to coax, prod, and bribe private employers into hiring more. That's better than nothing, but it means employers' preferences and concerns for profits are still determinative. Employment would still essentially function as a queue, with privileged workers at the front and marginalized Americans at the back. Khanna's program would cajole employers to move further down the line faster, but that's all.

The point of a Sanders-style job guarantee is to make life as difficult and unpleasant as possible for private employers: to make them increase pay and benefits whether they want to or not; to create permanent labor shortages, forcing business to hire all comers. Private employers couldn't be as picky about a worker's skills or education or background — or about their race or class or identity. That's a bold vision, but it also scares a lot of people; even many on the left. Khanna's program would conciliate and work around employers' power rather than directly challenge it.

What of the Full Employment Caucus' bill, HR 1000? It sits near the Sanders end of the spectrum. But it does contain some limitations, like a temporary freeze if unemployment drops below 4 percent or if inflation rises above 3 percent. Seeing as we're below 4 percent now, but U.S. employment still isn't delivering the benefits it should, that's an issue. HR 1000 also aims to answer the "how would we pay for it" question by establishing a trust fund. That might seem politically smart, but it could backfire: What happens to political support for the program if the trust fund runs out?

Then there's Booker's bill. It would start with a pilot program: Fifteen local areas in which a $15-an-hour-plus-benefits job guarantee would be tested for several years. The ultimate goal would ostensibly be a national version of the same, a la Sanders' bill. But it also offers a possible answer to the problem of Sanders' ambition: Start small, learn by trial and error, and then build the system out to a national level once we've worked out the kinks.

The downside here is one of political strategy. It's basically a bet that, once the pilot program has proven itself, the political opportunity will be there to pass the national version. But so far the moments when control of the White House and Congress align to make big transformative change possible have been few and far between. If you have the political opening to pass a job guarantee, should you really risk it on a test-run?

Let's assume Democrats take back the White House and Congress in 2020. Immediately playing for all the marbles on a job guarantee sounds intimidating. And there are certainly political and practical risks.

The thing is, not immediately going for all the marbles means you could still lose them all too.