Millions of Americans and their communities are shortchanged each year because the Census Bureau can't count straight and because many federal grant-in-aid formulas ignore a large population of "undocumented" Americans. This is a tragic result of the census undercount.

The undercount has plagued every decennial census since 1870. It is virtually impossible for the Census Bureau to count every inhabitant. In recent years, two factors have made the undercount discriminatory and punitive. One is the growing use of census population data in formulas to allocate federal funds. In fiscal 1975, for example, $33.7 billion was obligated for 75 formula-grant programs that used population or per-capita income data, in whole or in part, as a basis for dstributing money.This sum amounted to 22 percent of all state and local expenditures. Clearly, the distribution of these funds would change if there were a change in the population count.

Another factor is that the undercount affects some groups more than others. While almost two-thirds of the 5.3 million persons missed in 1970 were white, the rate of undercount was four times as high for blacks (7.7 percent) as for whites (1.9 percent). The rate for black males was 9.9 percent. The undercount among Hispanics, native Americans, and Asian Americans, in all probaility, was also disproportionately high. The undercount rate tends to be highest among blacks and whites who are already the most disadvantaged in the nation.

The undercount not only affects our social and economic well-being, but our political representation as well. Its impact on reapportionment is a complex constitutional issue that has not been resolved. However, the impact on fund allocations is a policy question for which there is an available policy remedy.

Thus, it is understandable that many elected officials, especially those from the cities with heavy concentrations of under-enumerated persons, complain that they are not getting their fair share of federal and state funds.The population data used in allocating those funds are based on the decennial census, which is admittedly inaccurate or incomplete. A Camden, N.J., official provided this perspective: "An undercount does [not] merely suggest that black Americans are being shortchanged, but the greater consequence is that all persons, black and white, poor and middle class, who live in cities like Camden . . . are being shortchanged."

Nearly everyone who examines this problem agrees that it is complex and freighted with inequities. Where then, might one look for solutions?

Even though the Census Bureau has taken extraordinary steps to reduce the 1980 undercount, most experts and the Census Bureau itself agree there will be an undercount in 1980. Clearly there is a need for a procedure to adjust the 1980 census count so as to avoid financial loss to millions of Americans and to hundreds of cities.The Department of Commerce and its Census Bureau have been urged on numerous occasions to develop and implement an adjustment procedure. Among those who have made such recommendations are elected officials, the U.S. Conference of Mayors, the Black, Hispanic and Asian American Advisory Committees of the Census Bureau, and the National Academy of Sciences.

The recommendations of the National Academy of Sciences are especially noteworthy. Last year, the academy was awarded a contract by the "department of Commerce to determine wether there are feasible methods of adjusting or correcting the census count to assure greater equity in the allocation of federal funds. The academy determined that "methods of adustment with tolerable accuracy are feasible" and recommended that the secretary of commerce direct the Census Bureau to develop and implement a plan to adjust the 1980 census count.

Secretary Kreps has yet to act on this recommendation. In a June 6 letter to Mayor Maynard Jackson of Atlanta, chairman of the U.S. Conference of Mayors census task force, Secretary Kreps expressed her views as follows:

"While the Panel on Decennial Census Plans of the National Academy of Sciences did recommend that such adjustments be made, it should be noted that the panel did not endorse a specific technique for accomplishing this task.

"As part of our continuing research, the bureau plans to convene during the spring of 1980 to review this matter."

The negative consequences of an undercount have been known since at least 1950.

Definitive studies have already been conducted. The academy was engaged to review those studies and to make a recommendation, which it did. Moreover, the proposed 1980 conference, at best, would offer relief in the 1990 census -- a decade away -- if at all. Certainly such a conference is no response to the NAS admonition that " . . . to meet the urgent demand for some greater degree of equity, a temporary expedient has to be considered." The NAS made it clear that it was talking about a "currently available" methodology to adjust the 1980 census.

The secretary's explanation that the academy report "did not endorse a specific technique for adjusting for under-enumeration" serves to cloud rather than to clarify the issue. The fact is that the Department of Commerce did not ask the academy to endorse a specific technique. Rather, it asked for and got a determination on feasibility. In fact, the academy cited three feasible adjustment techniques that could be used. The secretary may choose any one of them or initiate her own adjustment technique.

If the secretary does not act now to implement a "temporary," "expedient," "feasible" and "available" adjustment procedure, an awful lot of people, including needy low-income blacks and whites, will be short-changed for another decade. This is a matter of national concern.

If the secretary of commerce continues to delay, the president and the Congress should intervene before the 1980 census undercount and its Catch-22 consequences further erode the people's confidence in the federal government. CAPTION: Illustration, no caption, By Michael Crawford for The Washington Post