In recent weeks, several public figures have spoken of a new secret weapon for bringing down next year's budget deficit: the deflation dividend.

Because inflation has declined, the theory holds, federal costs will not rise as much as predicted next year; even in defense it will be possible to do everything the president wants for fewer dollars than experts had previously thought.

There is only one problem with this theory.

No such large dividend exists.

The main reason is that, while inflation this year is lower by 2 to 2.5 percentage points than even the administration projected, the likely rate for 1983 and 1984 is still about what the White House estimated a year ago. Its projections for those years were so optimistic that no large "dividend" is to be had.

For instance, the Congressional Budget Office said in a study this week that "price increases for defense purchases have slowed, but not much below rates anticipated in earlier administration and CBO forecasts."

Senate Majority Leader Howard H. Baker Jr. (R-Tenn.), after a meeting with President Reagan Thursday, said Congress would want to take advantage of the "deflation dividend" and slow defense spending without reducing the real rate of Reagan's planned defense build-up.

However, some Reagan administration budget officials said yesterday that any "savings" that may result purely as a consequence of lower than expected inflation "will be quite small." Generally, discussion of such a "savings" involve no more than $2 billion or so out of the $247 billion Pentagon budget projected for 1984, other sources said.

Inflation this year, as measured by the gross national product deflator, the economy's broadest price indicator, likely will be about 4.9 percent, well below the administration's 7.2 percent forecast published last January. But the same set of year-ago projections also showed a drop in inflation to 5.5 percent during calendar 1983 and to 4.9 percent during 1984.

Martin Feldstein, chairman of the president's Council of Economic Advisers, said this week he expects the deflator to increase about 5 percent in 1983. Current forecasts by most private economists show inflation running somewhat higher than Feldstein's figure or those earlier administration estimates.

Thus the differences between current inflation forecasts and those on which the Reagan budget numbers for 1983 and later years were based are not large, generally no more than about one-half a percentage point.

Moreover, a lower level of inflation does not necessarily translate into a lower deficit.

One reason is that a lower inflation rate tends in the short run at least to mean a reduction in revenues. Taxes are mainly based on incomes and incomes generally rise less in times of low inflation than they do when inflation is high. In the short run the loss of revenue when inflation falls tends to be greater than the gain on the spending side.

Those spending savings are still significant, particularly in the many large programs where spending is indexed or tied automatically to the inflation rate, as in Social Security.

But the falloff in inflation in 1982, to about 5.5 percent from about 8.5 percent, probably saved the Treasury only about $3 billion in programs that are indexed by law or by congressional custom--that is, where Congress tends to lift spending to keep pace with inflation every year anyway.

The longer inflation stays down, of course, the larger the savings. The inflation savings in fiscal 1983, the current year, for all programs except interest on the debt will be at least $20 billion.

Assuming that the inflation rate does not go back up, the decline in spending for the programs adjusted annually for inflation could be as much as $50 billion by 1984.

Even in some spending programs, however, a decline in inflation can mean an increase in the deficit. For example, part of the present decline in inflation is due to lower farm prices, and these have raised farm price support payments by several billion dollars.