When it comes to deficits, Congress is acting like an alcoholic trying to decide what to do with the key to the liquor cabinet. Last summer the Senate Republicans begged the president to hide the key. With the help of Tip O'Neill and some House Republicans, the president was persuaded to hand the key back to Congress, suggesting that the bottles labeled "military spending" and "Social Security" are okay to drink. This fall, in the Gramm-Rudman proposal, Congress is threatening to throw away the key so no one can find it.

The basic idea behind Gramm-Rudman, simply to mandate a balanced budget before 1991 by creating an artificial crisis, is appealing on an emotional level. But it is unappealing on an intellectual level. There are three major problems with the proposal: the target of a balanced budget is unnecessarily harsh; the recession-canceling mechanism won't work and would be unneccessary in a well-designed deficit-cutting program; and it lacks credibility because it resolves none of the basic conflicts between the president and Congress and within Congress that have prevented significant deficit reduction over the past five years.

There is a sound economic basis for replacing the Gramm-Rudman target of a balanced budget with a less harsh deficit target equal to about 2 percent of GNP. When deficits are large, at 5 percent of GNP as they have been since 1981, they cause debt to grow faster than GNP. Overall debt relative to income has risen faster since 1981 than during any peacetime period in history. To induce lenders to buy and hold more debt relative to their incomes, it has been necessary to offer higher real interest rates. Other undesirable corollaries to a rapid rise in the debt- to-GNP ratio have been a strong dollar, protectionist pressure, an inability to sustain adequate levels of domestic investment, and a rapid increase in interest on federal debt that could reach $400 billion per year by the early 1990s.

If the budget deficit could be cut to 2 percent of GNP, the ratio of debt to GNP would be stabilized, and that would remove a major source of upward pressure on interest rates. Based on historical norms, interest rates that today are running at about 12 percent, given a continuation of today's expectation of about 5 percent inflation, would drop to about 9 percent with a stable ratio of debt to GNP. A 3 percentage point drop in after-tax real interest rates would provide a natural stimulus to investment. The concurrent lower deficit would mean that less of our meager flow of saving would be absorbed by government dissaving. Less capital inflows would be required as less would be attracted at lower real interest rates. A weaker dollar and a lower trade deficit would also result.

Sen. Pete Domenici and his colleagues on the Senate Budget Committee proposed a three-year path to the "promised land" goal of a 2 percent deficit-to-GNP ratio early in this year's budget cycle, based on formulas that foresaw 3.5 to 4 percent growth. But 3 and 4 percent growth rates characterized the American economy during the 1970s and 1960s. The growth rate for 1985 will be about 2 percent, just the average rate for the 1980s. That lower rate is a more realistic planning device. The major implication for budget strategists is to slow the timetable for reaching Domenici's "promised land" to five years instead of three.

Recognizing that Gramm-Rudman could trigger sharp spending cuts during a recession, its authors propose that its deficit-cutting requirements be modified or negated if a recession is predicted. Economic forecasting, especially the constrained optimism of official forecasts that have never predicted a recession, is not up to the fine-tuning requirements of Gramm-Rudman.

Beyond its feasibility, the Gramm-Rudman recession-escape mechanism raises a fascinating question. If deficit reduction is the right medicine to restore the economy's vigor, why cancel the prescription in a recession when the economy is least vigorous? The metaphorical answer is that a weak patient can't stand strong medicine.

If the patient -- the economy -- may weaken unpredictably at some time during the course of treatment over the next five years, the only reliable safeguard is a milder prescription: aim, irrespective of the business cycle, for a 1991 deficit of 2 percent of GNP. This means abandoning discretionary fiscal policy, which hasn't been overly effective anyway, while we are bringing the deficit down to about $90 billion by 1991. The benefits of steady progress toward a stable debt-to-GNP ratio outweigh the risks of trying to fine- tune the economy during the transition.

The need is great for a credible five-year program of deficit reduction. The key word, however, is "credible" and that means designing a program that will last through political cycles as well as business cycles. Neither Gramm-Rudman nor any other formula will remove the need for the president and Congress to make changes they have so far refused to make. The Senate version exempts about half of total spending from cuts, while the House version exempts more than two-thirds. Both versions leave open the possibility of defense cuts or a tax increase, neither of which is acceptable to the president. If a December 1985 deficit forecast of $195 billion to $200 billion triggers an early 1986 sequester of about $30 billion, the result would be cuts of at least 10 percent in the one-third of spending left "on the table."

Faced with this conflict, Congress may be inclined toward tax increases on a scale that would spell failure for the president's policy of revenue starvation to force spending cuts. Alternatively, many in the House may be hoping that the Congressional Budget Office's role in the deficit-reduction process will be declared unconstitutional, thereby negating Gramm-Rudman. Having failed to deal with the deficit problem, the executive and legislative branches of our government now contemplate passing the buck to the judiciary branch.

Like the key to the liquor cabinet, discussion of budget procedure without attention to budget content is a red herring. Solving the deficit problem requires treatment of cause, not its symptoms. My prescription has three parts: 1) aim toward a deficit-to-GNP ratio of 2 percent by 1991; 2) resolve that if spending cuts alone are to achieve deficit reduction, then all spending must be cut; and 3) if all spending is not cut, raise taxes on consumption.