very serious drawbacks to doing it through the reconciliation process. It would require Obama to walk quite a political tightrope after dangling major healthcare reform in front of the voters:

But reconciliation is not just a “50-vote senate,” as it's sometimes called. It's a process constructed in the 1970s for a specific, limited purpose: to bring existing programs in line (reconcile them) with a long-term budget. Since then, it's been used for huge policy changes: the Reagan and Clinton budget plans, the Bush tax cuts. But there are limits. Under the senate's Byrd Rule, intended to hold the process somewhat to its original purpose, reconciliation can't include provisions that have no budgetary effect or that have an effect outside the current budget window, which right now is five years. (Byrd Rule limits can be waived, but by 60 votes, so you're back in the 60-vote Senate.)

To greatly oversimplify, what this means is that it's almost impossible to use reconciliation to build something new. You can expand Medicare or shrink it, cut taxes or raise them. But to construct something that doesn't already exist will inevitably require provisions that don't in themselves have a significant budgetary impact: regulations, structures, guidelines, realigned bureaucracies. In particular, much of the structure of health insurance exchanges that are envisioned in the House and HELP Committee bills would not survive the Byrd Rule axe. Only the flimsiest outlines of a health reform bill would survive – the financing would be there, but not the structures to ensure that the money would be used properly. Further, reconciliation would give the Finance Committee – which controls the money – even more clout over the more liberal HELP committee.

Some have suggested using reconciliation to install the rough skeleton of reform, and then fixing it later, but the act of using reconciliation in the first place is such a nuclear option that it is likely to poison the waters not just with the four semi-reasonable Republicans but also with the Democrats who are left out of the deal, and will be needed on subsequent legislation.

But what if Congress did it in reverse? Use the 60-vote Senate to pass whatever they can pass now -- we liberals will grumble but live with it -- and then use reconciliation next year to fix it. With the exchange structure and subsidies established, it wouldn't be hard to add an employer mandate, which would save money. With the rudiments of even a weak public plan in place, it wouldn't be complicated to expand it and modify its eligibility rules, in ways that might save or cost money but in either event, involve budget changes to an existing program rather than creating something new. Aggregating small changes over the next few years (on the model of the steady expansion of Medicaid engineered by Henry Waxman and others over the 1980s and 1990s) could non-controversially build the kind of robust and equitable system we dream of.

It's not ideal, and any political scheme based on do something now and hoping to fix it later faces the reality of all the partial reforms that litter the landscape. A plan that is so bad that it brings a backlash is more likely to be repealed than fixed. But it might just be that the big reform of health care can't be achieved all at once. And this would at least get the pieces in place for the next phase to move forward, with or without the current obstructionists.

The big problem there is if the Democrats are turned out of control in the next election. Then we have a very big problem on our hands.