General Motors should not be fed while still caught, because that will leave it trapped until we get tired of feeding it. Then it will die.

Michael E. Levine, a former airline executive and a distinguished research scholar and senior lecturer at NYU School of Law, writes in The Wall Street Journal:

General Motors is a once-great company caught in a web of relationships designed for another era. It should not be fed while still caught, because that will leave it trapped until we get tired of feeding it. Then it will die. The only possibility of saving it is to take the risk of cutting it free. In other words, GM should be allowed to go bankrupt.

Consider the costs of tackling GM's problems with some kind of bailout plan. After 42 years of eroding U.S. market share (from 53% to 20%) and countless announcements of "change," GM still has eight U.S. brands (Cadillac, Saab, Buick, Pontiac, GMC, Saturn, Chevrolet and Hummer). As for its more successful competitors, Toyota (19% market share) has three, and Honda (11%) has two.

GM has about 7,000 dealers. Toyota has fewer than 1,500. Honda has about 1,000. These fewer and larger dealers are better able to advertise, stock and service the cars they sell. GM knows it needs fewer brands and dealers, but the dealers are protected from termination by state laws. This makes eliminating them and the brands they sell very expensive. It would cost GM billions of dollars and many years to reduce the number of dealers it has to a number near Toyota's.

Foreign-owned manufacturers who build cars with American workers pay wages similar to GM's. But their expenses for benefits are a fraction of GM's. GM is contractually required to support thousands of workers in the UAW's "Jobs Bank" program, which guarantees nearly full wages and benefits for workers who lose their jobs due to automation or plant closure. It supports more retirees than current workers. It owns or leases enormous amounts of property for facilities it's not using and probably will never use again, and is obliged to support revenue bonds for municipalities that issued them to build these facilities. It has other contractual obligations such as health coverage for union retirees. All of these commitments drain its cash every month. Moreover, GM supports myriad suppliers and supports a huge infrastructure of firms and localities that depend on it. Many of them have contractual claims; they all have moral claims. They all want GM to be more or less what it is.

And therein lies the problem: The cost of terminating dealers is only a fraction of what it would cost to rebuild GM to become a company sized and marketed appropriately for its market share. Contracts would have to be bought out. The company would have to shed many of its fixed obligations. Some obligations will be impossible to cut by voluntary agreement. GM will run out of cash and out of time.

GM's solution is to ask the federal government for the cash that will allow it to do all of this piece by piece. But much of the cash will be thrown at unproductive commitments. And the sense of urgency that would enable GM to make choices painful to its management, its workers, its retirees, its suppliers and its localities will simply not be there if federal money is available. Like AIG, it will be back for more, and at the same time it will be telling us that it's doing a great job under difficult circumstances.

Federal law provides a way out of the web: reorganization under Chapter 11 of the bankruptcy code. If GM were told that no assistance would be available without a bankruptcy filing, all options would be put on the table. The web could be cut wherever it needed to be. State protection for dealers would disappear. Labor contracts could be renegotiated. Pension plans could be terminated, with existing pensions turned over to the Pension Benefit Guaranty Corp. (PBGC). Health benefits could be renegotiated. Mortgaged assets could be abandoned, so plants could be closed without being supported as idle hindrances on GM's viability. GM could be rebuilt as a company that had a chance to make vehicles people want and support itself on revenue. It wouldn't be easy but, unlike trying to bail out GM as it is, it wouldn't be impossible.

The social and political costs would be very large, but if GM fails after getting $50 billion or $100 billion in bailout money, it'll be just as large and there will be less money to soften the blow and even more blame to go around. The PBGC will probably need money to guarantee GM's pensions for its white- and blue-collar workers (pension support is capped at around $40,000 per year, so that won't help executives much). Unemployment insurance will have to be extended and offered to many people, perhaps millions if you include dealers, suppliers and communities dependent on GM as it exists now. A GM bankruptcy will make addressing health-care coverage more urgent, which is probably a good thing. It would require job-retraining money and community assistance to affected localities.

But unless we are willing to support GM as it is indefinitely, the downsizing and asset-shedding will have to come anyway. Even if it builds cars as attractive and environmentally responsible as those Honda and Toyota will be building, they won't be able to carry the weight of GM's past.

GM CEO Rick Wagoner says "bankruptcy is not an option." Critics of a bankruptcy say that GM won't be able to get the loans it will need to guarantee warranties, pay its operating losses while it restructures, and preserve customers' ability to finance purchases. While consumers buy tickets from bankrupt airlines, electronics from bankrupt retailers, and apartments from bankrupt builders, they say consumers won't buy cars from a bankrupt auto maker. But bankruptcy no longer means "liquidation" or "out of business" to a generation of consumers used to buying from firms in reorganization.

GM would guarantee warranty support with a segregated fund if necessary. And debtor-in-possession (DIP) financing -- loans that provide the near-term cash for reorganizing companies -- is very safe, because the DIP lender has priority over all other claimants. In normal markets, it would certainly be available to a GM that has assets to sell, including a viable overseas business. Such financing is probably available even now.

In any event, it would be lined up before a filing, not after, so any problems wouldn't be a surprise. As a last resort, we could at least consider a public DIP loan to support a reorganizing GM with a good chance to survive -- as opposed to subsidizing a GM slowly deflating.

The fate of Daewoo -- the Korean auto maker that collapsed in 2000 after filing for bankruptcy, leaving about 500 dealers stranded in the U.S. -- is often cited as "proof" that a GM bankruptcy won't work. But Daewoo was headquartered in a part of the world where bankruptcy still carries a major stigma and usually means liquidation. Daewoo's experience is largely irrelevant to a major U.S. company undergoing a well-publicized positive transformation, almost certainly under new management.

GM as it is cannot survive without long-term government life support. If it gets that support, it can't change enough and won't change fast enough. Contrary to Mr. Wagoner's brave declaration, bankruptcy is an option. In fact, it's the only option that merits public support and actually has a chance at succeeding.