Sen. Christopher Dodd, D-Conn., accompanied by Sen. Tom Harkin, D-Iowa, at a news conference on health care, Tuesday, June 23. (AP Photo/Harry Hamburg)

In the current battle over health reform, progressives may have set themselves up for trouble by pinning all their hopes on the creation of a government-run insurance plan. A public plan is not a bad idea -- indeed, it could be a critical element in successful reform -- but it could also easily turn out to serve the opposite purposes from the ones progressives intend.

All the proposals receiving serious consideration in Congress allow employers to continue to insure their workers and dependents directly. They also call for new "insurance exchanges" as an alternative means for individuals and employee groups to purchase coverage. If there is a new government-run plan, it would be one of the options in those exchanges.

The great danger is that the public plan could end up with a high-cost population in a system that fails to compensate adequately for those risks. Private insurers make money today in large part by avoiding people with high medical costs, and in a reformed system they'd love a public plan where they could dump the sick. Although the proposals before Congress aim to limit insurers' incentives to skim off the best risks, the measures are unlikely to eliminate those incentives entirely.

Entry into the public plan for the eligible employed would be a two-stage process. First, employers would choose between paying into the exchange and buying insurance directly to cover their workers. Unless the exchange is such a good deal that nearly all employers take it, firms with a young, healthy work force would tend to buy insurance on their own, while those with higher-cost employees would go into the exchange's pool. As a result, the pool would suffer "adverse selection" -- it would get stuck with a higher-risk population.

Second, within the exchange, the government-run plan would compete against private insurers, yet it would likely abstain from the marketing strategies used by private plans to avoid high-risk enrollees. This double jeopardy of adverse selection could then more than nullify the advantage the public plan derives from its lower overhead (as a result of less money going for salaries, profits, and marketing).

How should a public plan work? According to one model, the public plan would resemble traditional Medicare and have lower costs than private insurers by dint of its lower overhead and greater purchasing leverage, which would enable it to pay doctors and hospitals less. On that basis, it could underprice private plans and attract an immense enrollment (131 million people, according to one estimate).

Some supporters favor this approach because they see it as a step toward single-payer, which is exactly what the opponents fear. Squeezed by the public plan, providers might raise prices for patients insured by private plans, sending those plans into a death spiral.

But a Congress that is not about to adopt single-payer is unlikely to adopt a Trojan horse for single-payer. Some compromise proposals -- such as Sen. Charles Schumer's -- offer a second model, calling for a "level playing field" between private insurers and the public plan, including limits on the latter's ability to flex its purchasing muscle. But tight controls on its bargaining power might doom it entirely if it faces severe adverse selection.

Here's the delicate political problem: Depending on the rules, the entire system could tip one way or the other. Unconstrained, the public plan could drive private insurers out of business, setting off a political backlash not just from the industry but from much of the public. Over-constrained, the public plan could go into a death spiral itself as it becomes a dumping ground for high-risk enrollees, its rates rise, and it loses its appeal to the public at large. Creating a fair system of public-private competition -- giving the public plan just enough power to offset its likely higher risks -- wouldn't be easy even if it were up to neutral experts, which it isn't.

In any event, the success of a reformed insurance system is going to depend more on general features of its design, such as the rules that apply to all insurers and in particular whether premiums will be risk-adjusted (providing a bonus to plans with higher-risk enrollees and imposing a tax on other plans). A key question is whether the exchanges will serve nearly all employers, creating broadly shared risk, or remain on the margins as limited, high-risk pools.

There are a lot of ways to defeat reform, not just by blocking it entirely, but by setting it up for failure. Those who think a public plan is a good idea no matter how badly designed are not thinking ahead.