For years and years, the U.S. tax code has worked like a magic money machine.

Every year or two, Congress would make a fabulous gift to the American peopel - billions and billions of dollars. It waw called a "tax cut."

The newspapers printed big headlines. Everyone felt good. A grateful nation would reelect the Congress.

But, afterwards, thanks to inflation and real economic growth and some other factors, the taxpayers sent even more money back to Washington. This allowed Congress to make another huge "tax cut" a year or so later. It also allowed the federal government to grow bigger and bigger, devoting more and more money to its own programs, even as it shared a modest rebate with the taxpayers.

The taxpayers, unfortunately, don't really gain much from this magic machine. It has been pumping out periodic "tax cuts" for 10 years, winning political points for the operators, but the taxpayers wind up paying the same share of their incomes in federal income tax. Despite six major "tax cuts" since 1968, the effective federal income tax rate has stayed remarkably stable.

The actual average rate was 11.7 percent in 1969 and 12.5 percent last year. From 1960 to 1975, Congress enacted "tax cuts" worth 135 percent of what taxpayers were paying originally, an annual "savings" of more than $50 billion. But meanwhile, the actual revenue to the federal treasury from individual taxpayers tripled.

This year, the magic machine went a little haywire. The Republicans, tired of playing the scold on fiscal affairs, proposed a new game called Kemp-Roth. Instead of giving a one-year "tax cut" to the people, the Grand Old Party said it would be three times better to give them a three-year tax cut, all at once. This produced some breathtaking headlines about a "one-third tax cut" of $100 billion or more.

But if Congress gives away three years of tax cuts at once, what will it have left to give away next year or the year after?

Sen. Russell B. Long, the Finance Committee chairman who often explains taxes and politics with discomforting candor, raised that very question in debate.

"Senators should know when they are well off," he told his colleagues. "Houses members should know when they are well off. Here you have a situation in which, whether it is your fault or not, at least you will be able to vote for a tax cut. And if we cannot keep inflation from moving ahead, you can vote a tax cut again in the next Congress."

Democrats denounced the Republicans as irresponsible, then promptly countered with a "son of Kemp-Roth," which promised a 20 percent "tax cut" of $140 billion - spread over four years.

If Congress had not stopped this merry-go-round and killed off Kemp-Roth as it moved to adjournment, this might have continued until someone proposed a 100 years. To paraphrase the late Everett Dirksen, a billion dollars here and a billion dollars there and plenty soon you are taking real money.

This recent tax debate did revolve around some genuine economic issues. Should the federal government shrink a bit in relation to the national economy? If so, how do you do it/Neither of those questions or others were resolved, but Contress will be back next year to make new headlines.

Meanwhile, here is how the magic works: as inflation raises wages and salaries, people know it is an empty benefit, since everything coasts more to buy. But the U.S. tax code is progressive: people are supposed to pay a higher percentage as they make higher incomes. Thus, wage earners keep moving up to higher tax brackets - even though their additional dollars are each worth less. The periodic tax cuts, designed to offset that increase, really took hold as a habit in the last 10 years, as the nation entered a period of prolonged high inflation that began with the war in Vietnam.

At the same time, the national economy keeps growing in real terms. The population gets bigger each year and more products: more workers find jobs and pay taxes. All of this allows Uncle Sam to keep paring the income tax rates in the law, while confident that the nation will pay in more and more dollars, anyway.

The federal tax share of the gross national products has remained quite stable over the last 20 years, not withstanding all seasonal political rhetoric to the contrary. In 1958. federal taxes of every kind took 17.7 percent out of the economy; in 1977, they took 18.8 percent. Yet federal revenues and spending quadruppled over these years.

This year, the Carter administration and Congress were soberly debating whether the rebate should be $25 billion or $17 billion, when the Republcans grabbed center stage with their dazzling talk of $100 billion for the folks (or $78 billion or $98 billion or $112 billion, depending on whose estimate you believe).

This would have been a real tax cut. as opposed by an illusory one, if Congress had enacted it. But it wouldn't be nearly as radical as it sounded.

For one thing, federal revenues will keep growing especially if nobody figures out how to control inflation. According to one estimate by the Congressional Budget Office, with no changes in tax laws, federal revenues will grow by about $200 billion by 1981, the year when Kemp-Roth would have its final effect. So you might say Kemp-Roth and similar proposals would split the diffence, more or less with the government getting half and the taypayers half of the inflation-growth dividend.

The tax brackets on the tax returns would be trimmed in stages by roughly 35 percent, on average under Kemp-Roth. But the actual income tax rate - all of the income tax paid as a percentage of all income earned - would drop only modestly to about 10.5 percent . This is where it was in 1965 and it is only factionally less than the effective rate was in 1975.

Furthermore, what the federal government gives back with one hand, it takes with the other. The public as a whole comes out dead even, even if Kemp-Roth is enacted. This is because Congress has already scheduled substantial increases in the Social Security tax over the next few years.

Furthermore, what the federal government gives back with one hand, it takes with the other. The public as a whole comes out dead even, even if Kemp-Roth is enacted. This is because Congress has already scheduled substantial increases in the Social Security tax over the next few years.

By 1981, counting the "one-third tax cut" of the Kemp-Roth headlines plus the Social Security increases, which the politicians don't talk about much, the effective average tax rate would be 16 percent. That is right where is was from 1970 to 1975.

Of course, if the federal revenue is shifted from the progressive income tax to the flatter tax collected for Social Security, it tends to take more money out of some people's pockets and put it into somebody else's. Congress had been making the Social Security tax mildly more progressive for a decade, but the long-term effect of this tax is still to shift the burden toward the lower end of the economic ladder. On the other hand, the Democratic majorities in Congress, which habe enacted the periodic income tax cuts, usually give the biggest ones to the low end of the scale. Thus, wage earners above $20,000 complain that, when the government gives an inflation rebate every couple of years, most of the cash doesn't go back to the same people who sent it in.

With all these considerations, Senator William V. Roth (R-Del.), who coauthored the GOP plan, could make the claim on the Senate floor that this "one-third tax cut" would not take anything away from the federal treasury.

"I think it is important to recognize that the Roth-Kemp tax cut is no greater than ha anticipated through 1983," Roth said.

Even critics who disagree with him conceded, that, under certain economic conditions, namely, a week economy that needs tax cuts for stimulus, Kemp-Roth could be just the right medicine in 1981 or 1983.

The problem is that nobody knows what the economy will look like three years from now, let alone in five years. If the economy is booming, a huge tax cut could inflate everybody's dollars all over again. The experts make projections of all kinds, but even economists have limited faith in their predictions. Based on the economic forecasting of the last decade, their caution is fully justified.

That's where the political argument begins - and where it ended unresolved when Congress adjourned. Should the politicians try to look ahead and anticipate with tax cuts in advance? Or should they continue to enjoy the annual ritual of delivering billions back to their constituents?

In a theoretical world, an economist could argue: why not look ahead? Let Congress enact a huge tax cut, spread over three or four years, and hope that it works out as an economic stimulus or that future Congresses cut down spending to keep things in balance. If it doesn't work, Congress can always change the rates and raise them back to where they were.

But Congress dwells in the real world and the two things it finds politically most difficult to do are cutting spending and raising taxes. Opponents to Kemp-Roth and similar variations fear that future Congresses would not have the courage to raise taxes, even if the economic conditions demanded it.

In the happy era of the magic money machine, when Congress gave us six big tax reductions, only once did it raise taxes. That was Lyndon B. Johnson's war tax, better known as the 10 percent surcharge, which was needed to pay for Vietnam war deficits.

Everybody in Congress remembers what then happened to Lyndon Johnson.