IT'S THAT TIME of year again. This week, the president probably will decide the size of the raise federal employes will receive in October. Out of deference to a less than sympathetic public, he is expected to set the raise at somewhat below the 13.5 percent that government experts say is needed to keep federal salaries "comparable" to those of industry. But whatever figure he settles on will be illusory. It will not take into account the in-grade increases that most federal workers receive, the bonuses that executives will get, nor some whopping built-in benefits that are not included in the comparison with pay in the private sector.

It will certainly not address the central reality that federal pay levels if anything should be lowered, not raised.

Nevertheless, an orchestrated howl of protest, matching, in its poignancy, Lee A. Iococca's plea for alms for Chrysler, will rise from the federal employes' lobby. "This must stop," declared the president of one federal union last year, when Carter proposed to limit the raise, "before abject poverty becomes the mark of the average federal employe."

Any impression of civil servants being society's ragamuffins would be hard to reconcile with the data compiled in the July 1979 "Survey of Current Business," which shows that in 1978, federal employes earned an average of $18,988 compared with an all-industry composite average of $13,275.

Officials of the federal Office of Personnel Management (OPM) dismiss this comparison as an apples-and-oranges thing -- the apples being working Americans at large and the oranges the federal force, which has a much higher component of professional and administrative personnel.

But for those who were born earlier than yesterday, the explanation is a cop-out. Since 1962, the great increase in federal programming has not been in research and development, where professionals dominate, but in routine payment social programs which require hives of clerical worker bees. Yet, during that same period, the superiority of the federal pay average over the all-industry average steadily increased from 27 percent in 1962 to 44 percent in 1978.

Back when men wore watch chains across their vests, compensation for federal employment was something less than lavish. But by mid-century, government salaries had moved up substantially and, in 1962, Congress formally established the principle of "comparability" between federal jobs and equivalent work in the business world. The Bureau of Labor Statistics (BLS) was charged with designing annual surveys to provide information on salaries in the private sector for use as a yardstick against those in the federal service. Nobody apparently gave much thought to the fact that BLS is composed of government workers whose salaries would be directly affected by the results of the annual surveys -- and that these results would be worked into a recommended scale of pay increases by other government people in what is now OPM. This is not to suggest any dishonest number-switching on the part of the officials concerned; it is to suggest a built-in bias that has ample room for expression in the arcane method by which "comparable" federal pay formulas are determined.

"Comparability" does not mean that accountants, engineers, editors, secretaries and bottle-washers in government are paid on the basis of what their counterparts in industry earn. As one official of OPM admitted, "We are not trying to see that a secretary at the GS 7 level has the same salary as a secretary in a business office who does essentially the same work."

Instead, the government compares federal to private jobs using a medium called the federal "job standard." Job standards are lists of criteria and responsbilities that describe the various grades in the civil service.

The theory is simple: If the job standards describe the level of skill and responsibility of a GS 14, all BLS has to do is find the private sector employes whose jobs match that same standard. Things equal to the same thing will be equal to each other.

But they are not equal. On the private side of the equation, BLS goes out and visits employes, observing their actual duties. On the government side, what is matched to the job standards is that flight of the federal imagination known as the "job description." That is like equating the military knowledge of Clausewitz with that of the fellow in the loony bin who says he is Napoleon.

Relatively few government workers have the responsibilities, make the kind of decisions or exercise the latitude that their job description attribute to them. The writing of the job description is a practiced art -- and the higher the grade, the greater the scope for what one OPM official referred to as "journalistic freedom" in describing the alleged duties that justify a given level of pay. If you take what a job description says a federal manager does, and find someone in the private sector who does those things, you'll discover he has a high salary indeed. Although the Carter administration has proposed pay reform legislation, no "reform" will endow the comparability process with a semblance of reality unless it takes seriously into account this central fallacy on which the whole system now rests.

On top of this mirage, BLS has built an elaborate mechanism that further deprives the comparability process of any accountant-to-accountant simplicity. The federal version of comparability is based on something called the PATC survey -- an acronym denoting a selected group of "professional, administrative, technical and clerical" jobs in the private sector. The salaries of all the jobs at each level (as measured by the all-important "standards") are averaged together -- accountants, attorneys etc. at the GS 12 level going into one pot, those at the GS 11 level into another and so on.

On the basis of this series of statistical stews, BLS plots a "payline" for the private sector, running from GS 1 to GS 15. Invariably, when this payline is matched against the line that represents existing federal salaries, the private payline comes out running above the public one. The difference between the two determines the percentile increase OPM deems necessary to bring federal pay up to par.

The techniques used in this annual exercise, however, are seriously suspect. The process, according to BLS itself, is "surrounded by controversy because there is no single, generally accepted method of computing [the paylines]."

For starters, of the 436 white collar occupations listed in government work, only 22 are "matched" in the survey of the private sector. Some civil service occupations, such as that of IRS agents, customs inspectors and air traffic controllers, are excluded simply because they have no counterparts in industry.

More occupations, such as that of an editor, are excluded simply because the statisticians have deemed that a less extensive survey can still yield a statistically sound result.

But while the number of jobs surveyed may well be statistically adequate, when the number is as small as it is the selection of jobs to be surveyed can skew the private payline one way or another. The GAO, in a 1973 report on the survey, noted that the jobs selected for PATC "contained disproportionate numbers which were highly paid in the private sector, which resulted in an upward bias of the average work level rates." For example, although 70 percent of employes at Grade 5 were clerical, of the 10 jobs surveyed at that level in the private sector, only one was clerical -- which gave clerical work only a 10 percent influence on the outcome. Six of the survey jobs, on the other hand, were the higher-paying college graduate entry jobs; they represented less than 1 percent of Grade 5 employes, but had a 60 percent influence on that portion of the private payline. (In 1976, "weighted averages" were introduced into the process, belatedly offering a degree of correction.)

Similarly, at the loftier Grade 15 level the only three jobs selected for surveying in the private sector were (and still are) attorney, engineer and chemist -- occupations that are among the highest paid in private industry. The salaries business pays for people in those three fields is the basis on which all GS 15 salaries are annually determined to be lagging behind salaries in the outside world. (Since the payline above the GS 15 level is determined not by survey but by "extrapolation," a cynic might suspect that those at the top who choose the occupation to be surveyed are no dummies.)

Further distorting the private payline is a series of limitations placed on the survey's scope that effectively eliminates from consideration the salaries of three-fourths of all non-federal white collar workers.

By and large, the minimum-sized manufacturing or retail establishment in which BLS will seek pay statistics employs 250 people. That cuts out of the survey the salaries paid in 96 percent of all manufacturing plants and 98 percent of all retail shops. The upward pressure this selection exerts on the payline can be seen from figures put out by a different office of the Bureau of Labor Statistics, which show how average weekly wages increase in direct proportion to the size of the employing firm. The data show the average weekly wage in a firm employing 50 to 99 people to be 22 percent higher than for a firm with three or fewer employes, and in a firm employing more than 1,000 people the average wage is 38 percent higher than in one with 50 to 99 employes.

Also left out of the PATC survey are the employes of state and local governments, who have come to constitute 14 percent of the working population of the United States. (Their exclusion is mandated by law because of the disputable theory that their pay levels are not set by market forces). Federal salaries are substantially higher than those of any state except Alaska. An American Management Association survey puts the average salary of the mayor in cities of more than 200,000 population at $42,700, less than the average salary of the GS 15. The budget director in cities of that size averages $31,800 -- compared to the more than $50,000 salary of the chief budget officer in any federal agency. Piled on top of these are some important elements of compensation that the government doesn't even try to compare.

One factor simply ignored in the comparability process is an intangible -- job salary loss from any of a number of causes: layoffs due to localized or widespread recession, performance considered unsatisfactory by his superiors, bankruptcy of his company, etc. None of these threatens the federal worker to any significant degree -- surely a form of real, if hidden compensation. It doesn't however, show up on the comparative paylines.

Beyond the year-by-year security federal employment offers to the competent and incompetent alike is the ultimate security: the fact that for the federal employe there is no longer any mandatory retirement age -- he can go doddering into eternity at full pay, each year presenting him with a pay increase to bring him up to "comparability" with the private sector.

But if there is no legal requirement for the federal worker ever to retire, there is one hell of an incentive. The generosity of the benefits under the federal pension system is truly remarkable. The pension of an employe in the private sector is most generally calculated as a percentage of his average earnings during all the years he was employed by the firm. More generous plans base the pension amount on his average earnings of the last five years, which, because of both seniority and inflation, can be expected to be his peak earning years. The pension of the federal white collar worker is based on his highest three-year average.

In addition, federal pensions, unlike most private pensions, are indexed to the cost of living to afford the retiree complete protection against inflation. A federal employe who averaged $12,000 a year over his 30 years of service and retired five years ago, age 55, at an annuity of $20,000 (based on his "high three"), would see that annuity go above $30,000 this year. Assuming a modest 5 percent inflation rate, by the end of his actuarial life 18 years from now he would have collected more than $1 million in pension money.

Whether or not one considers federal pensions excessive, the importance of including these benefits in any valid pay comparison is hardly arguable.

The Carter administration's proposed "pay reform" legislation calls for pay comparability to be based on "total compensation." Estimates of the potential savings to the taxpayers range from the administration's "$3 billion over the next several years" to the Chamber of Commerce's "immediate savings of $10-$14 billion." A substantial part of the difference has to do with retirement benefits. For it is unlikely that the adminstration will delve very deeply into the true cost of pension benefits -- much of which is an unfunded liability. Even to make known the actual and projected costs would put the administration between an enraged public on one hand, and, on the other, a powerful federal employes' lobby, doggedly opposed to any linking of salary and benefits.

The notion that government pay is low and needs to be boosted if we are to attract quality people to serve is periodically given new support by presidential pay commissions. Typically, these commissions consist of prominent citizens (Nelson Rockefeller headed the last one) who appear to the public to be neutral and authoritative figures. They are, in fact, busy people, dependent on the career civil servants who do the commissions' real work and greatly influence the conclusions. Inevitably, the commissions recommend sharply increased pay, particularly for the elite class of civil servants who help shape the recommendations.

From time to time these recommendations are given credence in the public mind by newspaper accounts of a high government official who is returning to the presidency of the college he left, or to the executive board of his corporation, because he can't get along on his $50,000-$60,000 government salary. Unreported, because they are not newsworthy, are the multitudes of equally qualified applicants who scramble after the vacated job.

Although it is the exception rather than the rule, many government officials -- particularly at the cabinet and sub-cabinet lever -- are paid less than people carrying like responsibilities in business or industry. That is no national tragedy. The prospect of high pay is hardly the best motivation for public service. The assistant secretary who can't get along on his $60,000 salary may not be too well attuned to the public he is there to serve, and it's a fair bet that the public interest will not be mortally injured by his departure from government.

By any fair comparison the bulk of civil servants are far better compensated than the general public for whom they work. If comparability is to be achieved, it will require not an upward but a downward revision of government compensation. That's not likely to happen as long as the federal pay machine is designed and operated by the payees.