Administrators of company pay policy face three fundamental issues:(1) setting their companies’ overal pay levels in relation to thoseof other companies; (2) evaluating individual company jobs anddetermining pay relationships among them; and (3) determining payrelationships among individual workers within the same job. The last ofthese functions–and the subject of this article–is often accomplishedby establishing minimum and maximum pay rates for a given job orgrouping of comparable jobs, and providing for adjustments of individualworkers’ pay within this range of rates based on performance,seniority, or both. Special tabulations developed from the Bureau of Labor Statistics 1983 an 1984 national surveys of professional, administrative,technical, and clerical pay (PATC), which cover white-collar employeesin merdium and large establishments, show that: * Most white-collar workers are under rate range systems providingfor periodic merit (performance) reviews of their pay. * Sizable rate ranges are often established for individual companyjobs, especially at the higher professional and administrative levels. * In practice, however, differences between the highest and thelowest rates actually paid are generally much smaller than differencesbetween the maximum and the minimum rates specified for a range. Thedata base Information for this article comes from (1) internal work-sheetsprepared by BLS field staff in the 1983 survey to record job titles,formal rate ranges, durites, and responsibilities of company positionsmatching surveyed occupations and (2) answers to questions on pay plancharacteristic from the 1984 survey.
Approximately 3,100 establishmentswere studied in the 1983 PATC survey. For some 1,400 establishmentsproviding rate range data, the internal work-sheets contained theminimum and maximum pay rates for individual company jobs matching oneof the 101 occupational work levels in the survey. Each of these work levels, ranging from entry-level to managerialpositions, is covered by a written job description. Where several worklevels are surveyed within a single occupation, they are identified byRoman numerals–the higher the numeral, the greater the duties andresponsibilities. Each of the narrowly defined work levels representsfairly homogeneous work duties and responsibilities. Thus,classification of employees in accordance with these descriptionspermits summary and analysis of rate range characteristics for employeesperforming similar work, regardless of company job title or grade.
Exhibit 1 provides a hypothetical example of this job matchingprocess in a large headquarters establishment. In most cases, aone-to-one relationship exists between a company job and a PATC surveywork level; for example, only the company’s project engineer hasduties comparable to those in the PATC survey engineer V definition.Less frequently, one company job spans two PATC survey levels; someengineering associates better match PATC survey engineer I, while othersgenerally perform engineer II duties. Also, two company jobs atdifferent grade levels may at times equate to one PATC survey level, asin the case of the engineer and the nuclear engineer positions in theexample, which both match engineer III. For purposes of this study,matches similar to the engineer III illustration were excluded becausethey spanned more than one company rate range. These excludedsituations accounted for fewer than 10 percent of the 22,000 matches inestablishments reporting rate ranges.
The study focused on the width of company rate ranges–that is, thespread between minimum and maximum rates–and the relationship of actualsalaries to points within the ranges. In exhibit 1, the maximum rate is50 percent above the minimum rate in company grades 2 through 8, andslightly higher in grades 11 through 17. Such patterns, as found insurveyed establishments, will be discussed later in the article.
Respondents to the 1984 PATC survey answered the followingquestions separately for the professional-administrative and thetechnical-clerical worker groups; (1) What types of pay plans coveremployees in white-collar jobs? and (2) if workers are covered by rateranges, what boundaries are specified for the ranges; how frequently arerate ranges adjusted; what formal provisions, if any, cover normalhiring rates within rate ranges; and what point within the rate rangeequates to a job’s market value? Following is a description of thegeneral characteristics of rate ranges as revealed by the answers tothose questions. Rate range profiles Formal salary payment plans incorporating a range of rats for eachjob classification applied to about four-fifths of the white-collarworkers covered by the 1984 PATC survey. (See table 1.) In contrast,single rates for a given job–an important formal system for settingblue-collar pay–were virtually nonexistent for white-collar workers.Informal systems, which base salaries primarily on an individual’squalificiations, accounted for almost all of the remaining white-collarworkers. Informal plans covered about 5 percent of such workers in thelargest establishments (those employing at least 2,500 employees),compared with about one-fourth of those in establishments with fewerworkers.
With few exceptions, a minimum and maximum were specified for eachrate range reported. Within the range, an individual’s payincreases typically were based on periodic merit (performance) reviews.This approach covered more than four-fifths of the professional andadministrative workers and two-thirds of the technical-clerical groupwho were under rate ranges.
Pay progression for the remaining workersunder rate ranges either was automatic, determined by their length ofservice in the job, or depended on a combination of job tenure and meritratings. Rate ranges are typically adjusted once a year–a practice coveringabout four-fifths of each worker group studied. Less commonly,provision call for range changes at some other interval or on an ad hoc basis. After an upward adjustment in the rate range, some workers’rates fall below the new minimum.
Employers reported that such”subminimum” rates are usually raised at the employee’snext performance review or anniversary date. Most establishments pay new employees at a specified point orwithing a specified portion of the range. The 1984 PATC survey foundwide use of three distinct approaches, where by new hires were paid atthe range minimum, at some point between the minimum and the lowerfourth, or between the lower fourth and the middle of the range.
Eachapproach covered 20 to 25 percent of the professional-administrativeworker group. For the technical-clerical group, hiring at the minimumof the range pertained to 42 percent of the workers, and was at leasttwice as common as the other two hiring approaches. (See Table 1.
) The pace of advancement within a rate range is influenced in partby an employer’s perception of the market value of a job when fullyand competently performed. Three-fifths of the white-collar workerswere employed by establishments that regarded the midpoint of the raterange as representative of a job’s market value. These employersused the midpoint for controlling salary costs, that is, by filteringthrough that point only highly rated employees or the most experiencedemployees. About 15 percent of the workers were in establishments inwhich advancement would be expected to be faster because the midpointwas set below the market value of a job.
(It should be noted thatanother 15 to 20 percent of the workers were in establishments that didnot recognize this concept of a job’s market value.e Range width As mentioned earlier, rate ranges make it possible for individualsin the same job and establishment to be paid at different rates. The1983 PATC survey looked at the potential for such differences in theapproximately 1,400 establishments reporting rate range information.Although these establishments are not statistically representative ofthe full PATC survey scope, they do span all of its covered industriesand varying workd force size groups. Furthermore, the results areconsistent with findings from earlier Federal studies of salarystructure characteristics in the private sector. Employers generally agree on the basic rationale for rate ranges,but commonly vary the percent by which the maximum salary rate exceedsthe minimum salary rate in a range (its width).
Ideally, rate minimumsshould attract qualified job candidates while rate maximums should beset to reward and retain high achievers. In practice, however,employers see these as flexible boundaries that at times allow for ratesbelow the specified minimum, for hiring above the minimum rate, and forprogression beyond the maximum rate in the range. Thus, the prescribed width of the range may differ from the spread in rates actually paid. Among the PATC respondents, the maximum of a rate range mostcommonly exceeded the minimum by 50 percent, as shown in table 2.Nevertheless, many establishments had wider or narrower ranges.
For the89 survey work levels compared, the average spread ranged from 37percent for stenographers II to 57 percent for accountants V andattorneys V. In general, rate spread for professional-administrativejobs exceeded those for technical-clerical occupations. Few employers maintained a constant range width for all theirwhite-collar jobs. Among the 1,338 establishments reporting two or morerate ranges, more than four-fifths varied their range widths by at least5 percentage points, and differences of 20 percent or more were common.This largely reflects the tendency of companies to establish separatesalary schedules for major groups of white-collar jobs, such asprofessional-administrative and technical-clerical occupations. Asshown in table 3, the proportion of establishments with uniform rangewidths (a zero or 1-percentage-point difference between the widest andnarrowest widths) was much larger for similar types of jobs.Nevertheless, even within a grouping of professional-administrative ortechnical-clerical occupations, a majority of establishments had varyingrange widths. Actual salaries within rate ranges How widely do actual salaries vary within rate ranges? Are thereclusterings of salaries within ranges? To anwer these questions, actualsalaries were compared to several points in the corresponding rateranges–the minimum, the midpoint, and the maximum–and to the spreadbetween the minimum and maximum.
These comparisons, it must bestressed, were limited to salaries of workers in company jobs matchingPATC survey definitions; a company’s rate range for a labor gradenormally would cover a number of jobs, some within, and some excludedfrom, survey converage. As might be expected, clustering at or near the minimum of the raterange was most pronounced at the lowest work levels–the”entry” levels–of an occupation, where job skills aredeveloped in preparation for advancement to more responsible positions.The following tabulation illustrates this point by showing, for threeoccupations and two work levels, the percent of white-collar workerspaid within 10 percent of their rate range minimums: Because workers do not reamin in entry level positions for lengthyperiods, they normally do not advance far into their rate ranges.Conversely, because fully experienced workers are less often promoted tohigher work levels, they tend to be granted more within-grate wageadjustments.
Unlike the minimum rate, the midpoint of the rate range wastypically an establishment’s focal point for controlling overallsalary levels of company jobs. One measure of cost control used byemployers is the average salary of employees in a rate range expressedas a percent of the midpoint of the range. Values of about 100 or lessindicate that, on average, salary costs do not exceed theemployer’s market value of the job. Using this measure, 80 of the worklevels came in at 102 or less,while the remaining 9 topped out at 108. The latter comprisedexperienced drafters, engineering technicians, photographers,secretaries, and stenographers–groups that include many long-serviceworkers, some of whom were paid above the maximum of their rate ranges.not unexpectedly, some establishments allowed average salaries to risewell beyond the midpoint of the range.
Most establishments, however, paid only salaries falling within theassociated rate ranges. Moreover, it was common for substantialportions of these ranges to be unused at a given time, in part becauseos use of the midpoint as a salary control, or hiring at rates above theminimum, or both. To illustrate this point, the spread between thehighest and the lowest salaries actually paid was computed as a percentof the rate range spread for the job. On average, these ratios,indicating the proportion of the rate range being used, fell betweenone-third and one-half for professional-administrative work levelsstudies, and between two-fifths and two-thirds for techniccal-clericalclassifications.