Cutting taxes is to Congress what eating spinach is to Popeye: an invigorating pleasure scarcely to be resisted.

That being so, you might well blink at Congress behavior in the past six months. Citing the dangers of resurgent inflation, it apparently has rejected the joys of tax cutting for the less glamorous job of bringing the budget closer to balance. In January, the administration proposed a $25 billion tax cut for fiscal 1979. In August, the House passed a bill with $10.1 billion in fiscal 1979 reductions.

Few things are more unsettling than seeing a friend act totally out of character. The same truth appleies to institutions and, when Congress begins straying from well-established norms of behavior, you may legitimately wonder whether politics as you once understood it still exists.

Be assured: It does.

What Congress has done in the past nine months is to stage some masterful and engaging theater worthy of George M. Cohan. Although the final scenes of this song and dance are still playing, enough action already has occurred to make the outcome reasonably certain. When Congress completes the tax bill, the over-all tax reduction will be nearly as big as the White House's original proposal. Indeed, it's conceivable that it might even be slightly bigger.

Just how Congress has managed to do one thing - maintaining the size of the tax cut - while seeming to do another - cutting it down and reducing the deficit - is the subject of this short essay. We tell the story less in criticism than wonderment.

In a $2 trillion economy, the size of a tax cut, within a range of a few billions of dollars, does not alter the course of history. So no one can attach too much economic significane to s slightly bigger tax cut or a smaller deficit. But the story does illustrate the genius of politics at work: it's a magician's game in which all the pieces are moved about so fast that they create a blur, confusing the audience as to what actually happened.

One of the unlikely characters in this tale is the new chairman of the Federal Reserve Board, G. William Miller, who - wittingly or unwittingly - embrace a proposal that is critical to the entire deception. Hardly had Miller installed himself at the Fed before he was suggesting that as an anti-inflationary move, Congress might reduce the tax by postponing its effective date.

That Miller endorsed the proposal gave it more respectability than it deserves. For moving the effective date is simply a gimmick. The White House had proposed that the tax cut begin Oct. 1; Miller wanted Jan. 1. For most families, the three-month delay probably involves less than $100. But adopting this change - opened vast opportunities for tinkering and public confusion.

To understand this, you must remember that the government's fiscal year begins Oct. 1, nut the taxpayer's year begins Jan. 1. The budget deficit is always cited on a fiscal-year basis. In contrast, the effect of tax cuts, that is, how much money a "typical" family saves, is almost always calculated on a calendar-year basis. Shifting to Jan. 1 means that any given individual tax cut produces a smaller fiscal-year deficit.

You may not think that this makes much difference. No one could blame you for believing that postponing the tax cut for three months simply would reduce the revenue loss for the fiscal year by 25 percent, or put another way, that a $25 billion tax cut would cost about $19 billion in fiscal 1979.

This is as sensible as it is wrong. Tax collection is a complicated business, and one-twelfth of the tax revenues don't arrive at the Treasury each month. When a tax cut covers both the calendar-year and the fiscal year, the seasonal differences fade in importance; the revenue loss of the administration's proposal was about the same for both. But when the tax cut doesn't run the full fiscal year, the discrepancies can be large.

Thus did Congress reconcile the need for amply tax cuts with the need for a lower deficit. Remember that $10.1 billion cut passed by the House? That's $15 billion less than the original White House proposal (or 60 percent), but on a calendar-year basis, the $10.1 billion cut is worth $18.8 billion (about 25 percent less).

But that's not all. In addition to the major tax legislation, both the House and Senate are considering a variety of other small tax breaks. Among the possibilities: a tuition tax credit, credits for home insulation, and solar energy and tax relief for Americans working abroad. The House estimates the possible revenue loss of those measures at $1.6 billion on a fiscal-year basis. No one has yet estimated the calendar-year effect, but it wpuld be at least that, bringing the total tax cut (for the calendar year) to about $20 billion.

And, in fact, it is likely to be larger. The Senate and House regularly go through a ritual in which the House enacts lower tax cuts - allowing House members to advertise their "fiscal responsibility" - and the Senate enlarges the total, allowing everyone to take credit for being generous. The final tax measure may be $1 billion to $3 billion higher (for the calendar year) than the House's. Still, on a fiscal year basis, the cut is put at only $13,7 billion.

The actuall difference between this and the original White House proposal won't be worth whistling about. For, in fact, the actual reduction proposed by the Whie House was more like $24.3 billion, not $25 billion. The White House conveniently ignored (but Congress didn't) the expiration of a jobs tax credit for business: letting the credit expire increases revenues by about $700 million in fiscal 1979.

You might conclude from all this that Congress made losts of noise doing very little in the past nine months. That would be a mistake. In its details, the tax bill significiantly alters the original White House proposal, giving more of the benefits to the upper middle class and the wealthy. More important, the public relations struggle itself contains a monumental lesson: Caught between pressures for both lower taxes and lower deficits, Congress will try to satisfy both - at the expense, one suspects, of new spending programs.