DAVID STOCKMAN is at it again.

After he confessed to rigging the computers in 1981 to make the prospective Reagan deficits shrink, one would have thought that the budget director had had his fill of numbers juggling. But it was not to be. This time, of all things, Stockman has been fiddling with figures in the hope of demonstrating how fair the Reagan administration really is in its treatment of rich and poor.

Stockman unveiled his latest statistical wizardry before the Congressional Joint Economic Committee last month. He came fully equipped with charts, each to illustrate a remarkable assertion.

Claim 1: The poor have been affected only marginally by Reagan administration budget cuts. Indeed, if all of its proposed cuts for fiscal 1984 were enacted, Stockman contended, low-income benefit programs (food stamps, Medicaid, low income housing, child nutrition, Supplemental Security Income and Aid to Families with Dependent Children) would still be reduced only 5 percent below the levels sought by Jimmy Carter.

Claim 2: Large parts of programs for the needy weren't serving the poor anyway. Before the Reagan cuts, Stockman maintained, more than two-fifths of the benefits of low-income programs went to families with incomes exceeding 150 percent of the poverty line.

Claim 3: The wealthy really were not the big winners in the 1981 tax-cut act. In fact, he asserted, they had received less than 1 percent of the benefits.

It was an impressive performance, even if it was based on some peculiar evidence.

Start with Stockman's contention that actual spending for low-income benefit programs in fiscal 1982 and 1983 -- plus Reagan's proposed spending for fiscal 1984 -- is only 5 percent below the levels sought for these years in the last Carter budget.

Here Stockman has deftly made use of the high unemployment experienced under the Reagan administration in an effort to bolster his case. The costs of a number of these basic benefit programs vary with unemployment levels -- when more people are out of work, the number of households qualifying for the programs multiplies and program costs rise. By one estimate, for example, food stamp costs rise about $600 million for every percentage-point increase in the jobless rate.

The Reagan budget numbers Stockman cited reflect the impact of 10 percent unemployment on the costs of these programs. By contrast, the Carter budget numbers used by Stockman were calculated back in 1981, based on projections that unemployment would average only about 7 percent in the 1982-84 period. The result: Stockman was able to use the additional costs in the Reagan budget stemming from higher unemployment to make Reagan's spending levels look closer to Carter's -- thereby making the Reagan cuts appear smaller than they actully are.

Stockman was not content to stop his strange comparison there. Further manipulations occurred when he adjusted the Carter and Reagan budgets for inflation, converting both to "constant 1981 dollars."

To do this, Stockman adjusted downward both the actual Carter and the actual Reagan budget numbers for 1982, 1983 and 1984. He reduced the projected Carter spending levels for 1983 by 16 percent -- since the Carter budget had projected that prices in 1983 would be 16 percent higher than in 1981. And he adjusted the Reagan numbers for 1983 downward by just 10 percent -- the inflation level for 1981-1983 reflected in the Reagan budget. Since the Carter numbers were reduced by larger percentages than the Reagan numbers, this made Carter spending levels appear smaller in relation to Reagan's.

To be sure, such adjustments are valid in most cases -- but not for two of the major programs, Medicaid and subsidized housing.

Medicaid budgets are based on projections of inflation in health care costs rather than on projections of the overall inflation rate. When you do the proper inflation adjustment -- using health care costs rather than overall inflation -- you discover that the Reagan Medicaid cuts are about $2 billion deeper for the 1982-1984 period -- or more than double what Stockman indicated.

Equally egregious was Stockman's manipulation of the housing numbers. A substantial portion of federal outlays for subsidized housing consists of fixed costs under long-term contracts for construction or rehabilitation. These costs do not vary with inflation any more than a homeowner's fixed monthly mortgage payments do. Stockman had no business adjusting these fixed costs for inflation. However, Stockman adjusted these costs anyway, and reduced the fixed payments in the Carter housing budget by a greater percentage than he reduced the identical fixed payments in the Reagan budget.

This bit of legerdemain made it appear that Ronald Reagan -- whose administration has cut billions from new appropriations for subsidized housing, raised rents for all 3.5 million families and elderly persons living in subsidized units, and reduced the number of new low-income housing units being constructed or rehabilitated by more than half -- actually spent more on these programs over the past two years than Carter would have.

How significant are Stockman's manipulations? A new Congressional Budget Office analysis shows that as a result of the last two years of budget reductions, fiscal 1983 expenditures for the low income benefits programs were cut $5.2 billion below what they would have been had no changes been made by Congress. Stockman's chart, however, showed a reduction of only $1.7 billion. In other words, Stockman made two-thirds of the Reagan cuts disappear.

The administration's reductions, of course, would have been far deeper had all of its proposed cuts in aid to the poor been enacted. Of $20 billion requested last year in further cuts in these programs for the 1983-1985 period, Congress agreed to less than $4 billion.

Among the reductions rejected outright were administration proposals that would have doubled rents over several years for some of the poorest families living in subsidized housing, ended or reduced food stamps for more than 90 percent of the elderly who receive them, and sliced 700,000 low-income pregnant women and children from a food supplement program that has been proved to reduce infant mortality.

So much for Stockman Claim 1.

Next, Claim 2: that large chunks of benefits have been going to persons far above the poverty line. Indeed, Stockman maintains that before Ronald Reagan came to the rescue, average workers were being taxed to bring welfare families up to virtually the same standard of living as themselves.

Specifically, he contends that in 1981, 42 percent of all benefits in these programs went to families over 150 percent of the poverty line, and that 150 percent of the poverty line for a family of four that year was $13,390 -- or 92 percent of the median annual income for employed workers.

The misuse of statistics is particularly striking here. First, Stockman has compared 150 percent of the poverty line for a family of four ($13,390) to the median income for an individual worker. Sorry, but you can't do that. The real numbers go like so: The median income for a family of four in 1981 exceeded $26,000 -- not $13,390 -- and 150 percent of the poverty line is about half -- not 92 percent -- of the median income for a comparably sized family.

Then Stockman counted as part of the income of program beneficiaries the value of health insurance coverage provided by Medicaid and the benefits from living in subsidized housing -- but he did not include in his median income figures for workers either the comparable fringe benefits for employer- paid health insurance or the tax subsidies for mortgage and medical payments that many middle-income families receive. This makes for a neat comparison of apples and oranges.

When you do these comparisons properly, you find that the income and benefits of those partiWhen cipating in the federal programs were far below the living standards of average American families -- even before the Reagan budget cuts took effect.

Nor is Stockman's claim valid that 42 percent of low-income benefits went to families over 150 percent of the poverty line. While these figures are derived from Census data, Stockman misuses the evidence in ways that the Census Bureau itself warns against.

Drawing on the Census Bureau's work, Thomas C. Joe, a former Nixon administration welfare expert who now directs the Center for the Study of Social Policy, has prepared a devastating critique that shreds Stockman's claims on this issue. A few examples from Joe's analysis illustrate how deceptive Stockman has been.

A number of Stockman's "high-income" families were actually unemployed and receiving federal benefits for just a few months in 1981. Once back to work, they stopped receiving aid. But Stockman's figures reflect families' incomes for all of 1981 (rather than just for the months they actually received benefits), which enables Stockman to count many of these families as "high income" beneficiaries. The Census Bureau explicitly warns about this problem in the data, but Stockman ignored the admonition.

Similarly, the Stockman data distort income patterns when the composition of a household changes. The data attribute to households the income earned during the entire year by persons who were household members for only a small part of the year. Yet the absence of households members for part of the year (especially deserting fathers) may be the very reason that the remaining family members needed aid. The Census Bureau warns about this, too, stating that the data "may not always reflect the true economic status of the household during the year."

In short, the numbers Stockman uses have a major impact in exaggerating the number of high-income households receiving aid. When those distortions are removed, the picture is quite different. For example, Agriculture Department evidence that is free from these distortions shows that no more than three- tenths of 1 percent of food stamp benefits in 1981 went to families whose cash incomes exceeded 150 percent of the poverty line during the months they received food stamps.

While these manipulations are disturbing, Stockman's numbers-juggling reaches its zenith in his description of administration tax policies.

The wealthy, according to Stockman, received all of their tax cut when the top tax rate was lowered from 70 percent to 50 percent. Since this change constituted less than 1 percent of the tax benefits from the 1981 tax act, Stockman tells us, the wealthy ended up with less than 1 percent of the largesse and can hardly be described as the prime beneficiaries.

These startling conclusions contrast sharply with the findings of virtually every independent study of the 1981 tax act. The Joint Congressional Committee on Taxation, for example, found that the wealthiest 5 percent of taxpayers would gain 35 percent of the benefits from the tax act. Congressional Budget Office studies have shown that in fiscal 1982 through 1985, the tax and budget changes enacted under the Reagan administration will take more than $20 billion in benefits away from households with incomes below $10,000 a year -- while increasing the after-tax incomes of those making more than $80,000 a year by $64 billion.

How did Stockman get such different results?

First, his claim that lowering the top rate represented all of the tax cuts for the wealthy is nonsense. He simply ignores the plethora of new loopholes and expanded tax breaks incorporated into the 1981 act, such as the changes in estate taxes, IRA's and Keogh's, the All Savers Certificare and dividend reinvestment.

Second, Stockman carefully limited his definition of the wealthy (without informing his audience) to the top two-tenths of 1 percent of all taxpayers, those with incomes of more than $200,000 a year. This suited his purposes admirably: With so few taxpayers defined as wealthy, their aggregate tax benefits would not look so large. The sizable tax benefits going to the much larger number of taxpayers in the $50,000-$200,000 range were simply excluded from his calculations.

Moreover, Stockman omitted the fact that the elite group he did define as wealthy received, on average, a whopping $22,000 apiece just from the changes in tax rates -- before even counting the new tax shelter opportunities. Ronald Reagan himself saved $90,000 on his taxes last year because of the 1981 act. His after-tax income went up almost as much as if his salary had doubled.

The final part of Stockman's tax presentation was an attempt to discredit independent studies showing that those with high incomes received very large tax breaks.The problem with the studies, Stockman declared, was that for the wealthy, tax gains or losses stem less from rate changes (which the studies focused on) than from changes in the extent to which income is diverted into tax-free investments (which most of the studies did not treat).

Stockman's implication was that by reducing the top rates, tax shelters were being made less attractive -- and that declining use of shelters would reduce the gains for the wealthy below the levels cited in the studies.

While lowering the top rate may, by itself, reduce the use of shelters, Stockman again failed to disclose all of the facts: The 1981 act created so many new shelter opportunities that use of shelters has exploded despite the reduction in the top rate.

Use of syndicated shelters grew 12.5 percent in 1982, far more than the rate of inflation. Moreover, preliminary data indicate that for the first quarter of this year, syndicated tax shelter use is up 50 percent from the comparable period last year. The burgeoning use of shelters, which confer the preponderance of their benefits on the affluent, suggests that tax benefits for the wealthy from the 1981 act are likely to be larger -- not smaller -- than previous studies and analyses have indicated.

Finally, there is the question of purchasing power. A favorite Stockman (and White House) theme is that the average American's purchasing power has increased sharply during the Reagan presidency because inflation has come down so much. Purchasing power, however, is not determined solely by prices -- but by the interaction of prices and wages.

What the administration has failed to say is that wages have come down about as much as prices, leaving the average American with virtually no gain in real purchasing power.

What is widely regarded as the best measure of purchasing power -- the Commerce Department's index of real per capita personal disposable income -- shows that purchasing power under Reagan has increased at an annual rate of only four-tenths of 1 percent. This is well below the average rate of increase under every other president for the past 30 years.

The growth in purchasing power that did take place in the Reagan years occurred primarily from January 1981 until August 1981 -- before the Reagan economic program took effect. Since August 1981, when the Reagan tax and budget program was enacted, real per capita personal disposal income has declined. The average American's standard of living has fallen since the Reagan administration's program was enacted.

Stockman's manipulation of the numbers makes rational debate on these spending issues more difficult. But perhaps most significant is the new dimension that Stockman has added to the much-discussed "fairness" issue.

For what can raise more basic questions about whether this administration is fair than when one of its principal officials -- with access to data, staff and resources that few others in this town possess -- utilizes this power to rig the terms of the debate and misrepresent the nature of his administration's policies.