Once again Congress and the president are on a collision course over taxes and spending that threatens to prevent any future progress in shrinking the budget deficit. Despite the president's frequently stated opposition to any tax increase, the recent congressional budget resolution instructs the tax-writing committees to come up with $19 billion of new tax revenue. A novel feature of this year's resolution, designed to put the president on the spot and increase the likelihood that he will accept a tax hike, is an explicit link that reduces the level of defense spending if the president rejects the tax rise.

The president has nevertheless responded by reiterating that he will not raise taxes. His natural aversion to taxes is reinforced by some simple political arithmetic. For the president and his advisers, the political costs of accepting a tax increase far outweigh the potential benefits of a small deficit reduction and an even smaller increase in defense spending.

Major national opinion polls show that half of the public thinks that the deficit can be reduced without raising taxes and a majority opposes a tax increase at this time. The polls also show that the public credits the president and the Republican Party with the major tax cuts of this decade and that an overwhelming majority believes that the Democrats are much more likely than the Republicans to raise future taxes. This reputation is too valuable a political asset to throw away as the 1988 election approaches in order to raise defense spending by 2 percent and to cut the budget deficit by 5 percent or less -- especially since both the defense increase and the deficit reduction could easily be reversed by a subsequent Congress.

By accepting an explicit tax hike the president would not only deny the Republicans a major advantage in the 1988 campaign but would give the Democrats a major campaign issue. The Democrats could attack the president for failing to honor his 1984 promise not to raise taxes and could say that they and Walter Mondale had been right all along when they claimed that the president would have to increase taxes. Although the net cut in taxes during the eight years of the Reagan presidency would still have been unprecedentedly large, any tax increase now would seem to many in the public to be a return to higher taxes. The Reagan legacy of tax reduction, one of the achievements in which the president takes great pride, would undoubtedly be tarnished.

Despite this bleak political arithmetic, there is a long-shot political compromise on deficit reduction that might just be possible. There are two key ingredients in such a deal. First, it must raise real revenue without increasing income tax rates or even the excise tax rates on tobacco, alcohol or gasoline. The president and Congress must be able to say that they did not raise taxes. Second, the compromise plan must reassure the president that the additional revenue will be matched by equal spending cuts. In 1982, the president supported a congressionally proposed tax increase in exchange for what he believed was a congressional promise of even larger spending cuts. Since those spending cuts did not occur, the president has been distrustful of congressional promises to find spending cuts to match the increased revenue. What's needed, therefore, is a simple and unambiguous way to cut spending at the same time that the revenue increase is enacted.

The best way to meet both requirements would be to modify indexing, the automatic adjustment for inflation, on both the tax and the spending sides of the budget. Under the existing rules of indexing, tax brackets are automatically adjusted for any inflation that has occurred during the past year. If automatic tax adjustments were instead limited to the amount by which inflation exceeded some threshold level, there would be additional tax revenue without disturbing the existing consensus on tax rates.

In practice, a 3 percent threshold would mean a 1 percent indexing adjustment if inflation is 4 percent and a 2 percent adjustment if inflation is 5 percent. Individuals would therefore still get an automatic tax break every year to offset the inflationary bracket creep. Taxpayers would remain fully protected against the big tax hikes produced by rapid inflation.

Although a 3 percent floor on tax-bracket indexing would yield only about $7 billion of additional revenue in the first year, this would rise to about $25 billion a year by the third year. A very important feature of the inflation-indexing modification is that it would raise gradually more and more revenue in each future year and would do so without any explicit increase in taxes.

An essential feature of the plan to modify indexing would be to extend it to transfer programs such as Social Security and other government retirement benefits that are not based on individual need or low income. If the same 3 percent floor were applied to these programs, outlays would fall by about $7 billion in 1988 and $25 billion after three years -- a spending cut equal to the rise in tax revenue. The indexing modification would slow the rise in benefits but would not reduce those benefits or the protection that indexing currently provides against the adverse effects of rapid inflation.

A particular advantage of the indexing modification is that it would automatically link the increase in revenue and the decrease in outlays in a single legislative change that would reassure both the president and members of Congress that spending cuts and outlay increases would be changed by equal amounts.

The Ways and Means Committee is now beginning a futile search for a package of tax increases to raise $19 billion. It would be far better if the committee, which has jurisdiction for Social Security benefits as well as taxes, shifted its focus to an indexing modification that would make a powerful contribution to deficit reduction without explicitly raising taxes or reducing benefits.

Martin Feldstein was chairman of the Council of Economic Advisers. Kathleen Feldstein is an economist.