Ramesh Ponnuru makes a compelling case that “universal” health insurance coverage should not be the primary goal of health-care reform. The case is simple: the cost shifting which politicians point to as a reason for mandating coverage for everyone is relatively small. It can be addressed in less destructive ways and is a less critical problem than other issues (e.g. soaring costs and lack of portability).

But there is another reason to avoid “universal” plans: so far they haven’t worked. In Massachusetts, the best effort to use “market” reforms to attain universal coverage has resulted in less than universal coverage and rising costs. California abandoned its proposal and Hawaii pulled the plug on its universal coverage system.

Once you force coverage for everyone the cost skyrockets — unless of course you start limiting care, cutting doctors’ fees, and controlling what sort of treatment you will pay for. That has been the experience in the western countries that have gone down this road. The alternative — to try to stimulate competition among insurance carriers, have individuals purchase and control their own insurance and provide subsidies for the hard-to-insure cases — seems infinitely more attractive than duplicating the problems of other countries.

But the Massachusetts plan has been edifying, and perhaps the argument is better made by allowing fifty separate state experiments to bloom rather than creating a one-size fits all federal system. Maybe one of those fifty, or more than one, will hit on a formula for expanding coverage, improving care and reducing cost. The odds are certainly better than  they would be under a single untested federal plan. So if “something must be done,” a system of block grants and a period of experimentation might be in order.

That would be less grandiose than a single federal scheme. But it might do less damage and produce some real world experience on which we can draw to create a workable national system.