Why ObamaCare's Cost Controls Will Fail


Before the passage of last year's health care overhaul, supporters of the law repeatedly made the case that it was intended to help control costs. Obama and various administration officials promised that it would hold down health insurance premiums while restraining the long-term growth of health care spending. They rigged the bill to get the Congressional Budget Office to score the law as a net reduction of the deficit. 

But despite all the effort put into gaming the numbers, the fiscal argument was never very believable. As in Massachusetts, the bill was about coverage expansion first, cost-control second. And the law fundamental approach to cost-control was to hand off responsibility to bureaucrats, who would then look for successful cost-control innovations and try to replicate it throughout the system.

It was an belief born out of years of legislative failures: Congress has failed to control costs. Maybe, thought reformers, empowered experts can do it better. It's reform built on faith in bureaucrats, experts, technocrats, and policy wonkery. That's how we got the Independent Payment Advisory Board. And it's how we got the law's 400-plus page regulation for accountable care organizations (ACOs)—highly integrated provider networks that, in theory, have financial incentives to provide better, cheaper care. 

At best, these sorts of centrally planned reforms are untested, with no guarantee that they'll work—if anything, there's evidence suggesting that they might not. But the Obama administration and its health policy technocrats are still wedded to bureauwonkism anyway. Here, for example, is their latest round of self-congratulation regarding ACOs, via The Hill:

The agency said it has seen strong results from a five-year demonstration project with goals that are similar to ACOs' — lowering costs by improving quality and shifting away from paying doctors to perform more procedures.

The demonstration program involved 10 large, integrated healthcare systems. Seven of the 10 met all 32 of the program's quality benchmarks, the Medicare agency said in a release. And all 10 agreed to participate in a two-year supplement to the initial demonstration project.

But don't think the wonks have won this round: 

Most of the organizations that took part in the demonstration project, however, have voiced serious concerns about the proposed structure of ACOs.

Nine of the 10 health systems involved in the demonstration signed a letter in May saying they might not participate in the ACO program unless the Medicare agency makes major changes to its initial proposal.

The clinics said they all supported the concept of integrated, coordinated care, but that, "as currently proposed, ACOs have a greater potential for incurring losses … than for generating savings."

Reformers are mostly right when they argue that Congress hasn't done a great job of controlling health costs or the growth of tax-financed health spending. But where they go wrong is in thinking that a small number of experts in positions of government-granted authority will somehow be more successful. The basic problem with the bureauwonk model of health reform is that it assumes that technocrats can not only identify but successfully scale local innovation to the national level. That turns out to be exceedingly difficult; what works in one health organization doesn't always work in others. Even organizations willing to participate in demo projects won't always want to turn control of their innovations over to federal regulators. 

When CBO director Douglas Elmendorf testified before Congress last month about the government's history of attempts at health care innovation, he was blunt about the general lack of success: "The demonstration projects that Medicare has done in this and other areas are often disappointing…It turns out to be pretty hard to take ideas that seem to work in certain contexts and proliferate that throughout the health care system. The results are discouraging." The Obama administration is betting on these reforms anyway; a better bet might be that the results, or lack thereof, will continue to be more or less the same.