I have lost count of how many columns I have written denouncing Amtrak as a colossal waste of money. A half-dozen? A dozen? Probably somewhere in between. But Amtrak is still chugging along, and I am still railing against it because it so aptly symbolizes a much wider problem: our inability to end -- or modify -- old programs that have outlived their usefulness, if they ever were useful. This is a root cause of our budget problems and a larger failure to reassess the role of government.

Government shouldn't be an instrument for anything that, at some point, seems worth doing. But that is why we have Amtrak and countless other programs. They serve no genuine national need. Worse, we seem incable of doing anything about this. In politics, inertia reigns. What exists tends to survive. We have what writer Jonathan Rauch calls "demosclerosis" in a forthcoming book by the same name. Government becomes committed to past policies; it suffers a "progressive loss of the ability to adapt."

Until recently, Amtrak contended it was gradually reducing its dependence on federal subsidies. Given deficit consciousness, this was good public relations; it was also an illusion. For fiscal 1995, the Clinton administration is asking for a 9 percent increase in Amtrak's subsidies to $988 million, and the General Accounting Office says that larger amounts will be needed to prevent a decline in the amount or quality of service. The average age of Amtrak's passenger cars is 23 years; new purchases will be required. Passenger trains simply aren't self-supporting in a country where housing and business are so dispersed.

Since 1971, Amtrak's subsidies exceed $16 billion. Assuming all that was borrowed, these subsidies cost about $1 billion in annual federal interest payments (that's $16 billion X 7.2 percent, the average interest rate on federal debt). And for what? Amtrak provides only 0.7 percent of intercity transportation. Its annual ridership of 22 million has barely increased since 1979. By contrast, the number of airline passengers rose by 149 million (to 441 million) over the same period. Amtrak provides few energy conservation or anti-pollution benefits, because it is such a tiny part of the transportation system.

Our political inertia is not simply a byproduct of the so-called "iron triangle" of interest groups, executive-branch agencies and congressional committees that protect their favorite programs. The larger cause is an attitude that Americans bring to government. Despite rhetoric about "waste," we believe that government benefits, once bestowed, shouldn't be withdrawn. Ordinary Americans (the reasoning goes) depend on this or that program, and revoking it would be "unfair." The "iron triangles" become powerful because they have no genuine opposition. What's lost in this twisted, though understandable, logic is whether programs serve anyone but their recipients.

Naturally, beneficiaries argue that their programs serve a broad national interest. The political system is supposed to "determine whose claim is in the public interest and whose isn't," as Rauch puts it. Unfortunately, the system has defaulted. It no longer asks, debates and answers these basic issues. Of course, Amtrak claims it provides vital services. But almost everyone who travels on Amtrak would go by car, bus or plane if there were no Amtrak. Government can't do everything for everybody, and we have to distinguish between genuine national interests and publicly financed conveniences or giveaways.

Dozens of programs (farm subsidies, art and cultural subsidies, regional economic subsidies) would have a hard time surviving any rigorous assault on their public worth. If there were no farm subsidies, farmers would still produce and we would still eat. What exactly is the national interest? But there is no rigorous assault. For instance, Vice President Gore's report on "reinventing government" studiously avoided questioning individual programs. Instead, Gore concentrated on more opaque issues such as procurement and personnel policies, where everyone can agree that government ought to be more efficient.

The result is that, contrary to claims by the White House, there has been little overall discipline of domestic spending. Budget deficits are dropping mainly because defense spending is declining and taxes are rising. Consider. Between 1993 and 1999, the White House expects the deficit will fall from 4 percent of Gross Domestic Product to 2.1 percent of GDP. Over the same period, defense spending is projected to decline by 1.7 percent of GDP and taxes rise by 0.8 percent of GDP. This more than accounts for all the deficit reduction, even ignoring interest savings from less borrowing.

At some point, government spending, taxes and deficits overburden the economy, exceed the public's tolerance, or both. All the pressures will intensify in the next century when the "baby boom" reaches old age. Even now, a third of federal spending already goes to support older Americans; this will rise. "As an economic matter, fiscal responsibility will almost never be easier to attain than in the current period," write E. Eugene Steuerle and Jon Bakija of the Urban Institute. We should be using (but aren't) the 1990s to prepare for the 21st century by paring back government and limiting its responsibilities.

If we can't get rid of Amtrak, farm subsidies, art subsidies and similar programs -- programs that serve no national need and have relatively small constituencies -- then how can we ask harder questions? For instance, when Social Security and Medicare were created, most older Americans were considered poor. Now they aren't. Should government, then, be less generous to the affluent elderly? Because we live longer, should benefits begin later? What's being lost is an opportunity to improve the democratic conversation by asking basic questions about government's role.

Government needs to adapt to change, as Rauch warns. Trains were once an economic and socially useful way of moving people. But the car, the plane and suburbanization effectively eliminated their role. Amtrak defines the dilemma of democratic government. Programs tied to the past will fail the future.