fiscal year 1961. With the reorganization of the Department of the Army in 1962, the number was reduced to 23.

The private sector of the sample consisted of 23 prime contractors' laboratories, each selected because its activity was similar to one of the Army laboratories in the survey. In some cases, the "major" contract for a private laboratory was quite small but was included in order to encompass at least one private contractor engaged in the same type of R&D effort as each Army laboratory. To maintain pricing continuity, each respondent in the private sector was requested to report expenditures and prices on all contracts held for a designated technical service. The reporting unit in each instance (even for large firms having several laboratories) was the laboratory unit which had responsibility for the contract under study.

The information on fiscal year 1961 expenditures provided by the sample laboratories was used to determine the categories of items to price and to weight the indexes of prices. Major cost categories were then selected by BLS and a pricing survey

was initiated for 1961-63. Under general guidance of BLS personnel, respondents selected the specific items to be priced within a major category and provided specifications and prices for each of them.

Index Construction. The indexes were constructed by using a modification of the Laspeyres formula; i.e., annual price indexes weighted by expenditures during the base period (fiscal year 1961).*

Weights. To obtain weights for direct labor costs, the Army laboratories provided payroll data (for the 2-week period ending June 30, 1961) broken down into four classes-scientists and engineers, technicians, craftsmen, and unskilled employees. Each laboratory also reported the number of persons employed and the salary or wage paid for their most common jobs directly charged to R&D projects as of June 30, 1961.

Information on materials, supplies, and equipment provided by Army laboratories included (1) total obligations, (2) a listing by 2-digit code of sufficient Federal Supply Classification groups to account for at least 50 percent of total obligations, and (3) a further listing of items purchased to

account for at least half of the expenditures in each of these FSC 2-digit groups.

The Army laboratories also provided total overhead cost obligations chargeable directly (1) to R&D and (2) to general and administrative support. A further breakdown into costs of approximately 25 categories of these obligations was requested, and though most of the laboratories could furnish detailed lists of support costs, others (especially those receiving "reimbursable" support) could not.

Private laboratories provided total (fiscal 1961) costs chargeable to the contract under study subdivided into direct labor costs, materials costs, other direct charges, overhead costs, and fixed fees. Separate payroll data were furnished for scientists and engineers, technicians, craftsmen, and unskilled employees for the pay period ending nearest to June 15, 1961.

Each respondent was asked to list a sufficient number of items, or classes of items, to account for at least 50 percent of his total 1961 costs for materials, supplies, and equipment chargeable to the contracts. Fiscal 1961 costs were also requested on a list of 15 overhead and other cost items, or enough items to equal at least 75 percent of the total spent in these categories.

It was assumed that the laboratories surveyed were representative of all private laboratories holding Army R&D contracts, and therefore the indexes for private laboratories and for Army laboratories were combined by weights representing total obligations to all laboratories rather than totals only for laboratories in the sample. Data on

where the (-1) is the index for the preceding period, and the (P-1)'s are prices in the preceding period. As items become less important or as new items are introduced, substitute specifications are used in the index calculation, and the formula is not a precise representation of the calculation method. The price relative P/Pi-1 becomes P/P-1, where the P and the P-1 are prices of the substitute item in the current and preceding periods, respectively. That part of the price change due to the specification substitution is therefore not reflected in the index movement.

5 It was assumed that 26 times the sum of these amounts would equal the total fiscal year 1961 direct labor costs for Army laboratories. A check indicates the assumption to be correct within plus or minus 5 percent.

Army obligations to private contractors were distributed between labor, material, and overhead on the basis of expenditure patterns obtained from the schedules.

Pricing Methods. The 1961 expenditure survey provided the basis for the selection of representative categories of items for pricing, but accurate price indexes require continuous price data. Therefore, prices were collected for purchases and payrolls in July 1961, 1962, and 1963, for a set of carefully defined products and jobs whose specifications (including quantities, delivery terms, terms of sale, etc.) or descriptions did not change over time.

The Army laboratories selected from among professional personnel their two most common General Schedule (GS) grades and the most common GS grade for new employees. The most common annual salary for each grade as of June 30, 1961, and the most common job title for each classification were then reported. The three most common occupations, GS grades, and salaries for technicians were selected, and for wage board employees, the hourly rates and three most common job titles for craftsmen and the two most common job titles for unskilled workers were selected.

Because administrative increases involve no change in job description and because it is not necessary for an employee to acquire new skills to qualify for them, administrative increases or ingrade promotions were considered to be price increases and not quality changes. Grade raises and decreases, on the other hand, were treated as quality or specification changes and not price changes. Pricing for each year was based on the salary or wage rate paid for the type of work performed for each job.

The procedure for measuring changes in direct labor costs in the private sector was essentially the same as that used for the Army laboratories. Pri

Current legislation which requires that an employee's work be of "an acceptable level of competence" in order to qualify for an in-grade increase and the time requirement for in-grade promotions could, however, lend weight to an argument for treating in-grade increases as quality changes and not as true "price" increases.

For the private contractors, no adequate job descriptions were available to determine whether the salaries and hourly rates reported from year to year were for the same work as had been reported for a given job title in the base period. Reliance had to be placed on the likelihood that the same job title used by the same reporter in successive years described the same job duties. It was not necessary, of course, that the same job title used by different reporters describe the same job.

vate contractors were asked to describe their three most common types of scientific and engineering personnel according to work level, degree of responsibility, and annual salary as of June 30, 1961. The most common job titles for technicians, craftsmen, and unskilled employees were also selected by the respondents, but without regard to work level or degree of responsibility."

For pricing materials, each Army or private respondent was asked to select several specific products representative of the major cost categories listed in his base period report. Cost categories were selected by BLS and the respondent provided both specifications and the purchase price for a representative product within each category.

All 23 Army laboratories furnished prices on materials; the number of prices furnished ranged from a minimum of 2 items to a maximum of 31. Altogether, the Army laboratories furnished prices of 289 items for the 3 years of the study; only 1 item was priced by more than one laboratory.

It was not feasible to price all of the materials and support cost items listed by each respondent, and in some cases, the price movements for priced items of a laboratory were imputed to movements of nonpriced items purchased by the same laboratory. In cases where a laboratory had nonpriced items for which no similar products were priced, the price index developed for these categories in other laboratories was used, but in no instance was an index from the private sector used as an esti mate for the Army sector, or vice versa.

The pricing method employed for support costs in the Army laboratories and for overhead costs and other direct charges in the private laboratories was essentially the same as that used for materials and personnel. BLS selected major cost categories from each respondent's expenditure report and requested the reporter to select a specific segment of each category for pricing.

Problems. Though pricing overhead was easier than anticipated, detailed weighting data on overhead and other direct costs were difficult to obtain from private laboratories. In the base-period reports, only seven companies were able to provide a breakdown of their overhead expenditures. In some cases, a contract allowed a percentage of direct labor costs as an overhead charge, and the company was not required to maintain detailed

records. In other cases, overhead for the reporting unit was estimated as a percentage of a corporate pool, and, again, no detailed record was kept. The difficulties in obtaining satisfactory weight data on overhead from the private laboratories also affected pricing. To the extent that detailed expenditure data were unavailable, the sample of items was weak.8

Pricing Problems. Materials priced for the Army laboratories were nearly all "shelf" items, i.e., items purchased each year in some volume and under the same or similar specifications. Some of these were new to the market and care had to be exercised not to introduce them into the index too early, i.e., when they come out of the development stage and their relative price is decreasing rapidly. If a laboratory purchased an item only once during the period studied, an index was computed from price data obtained from vendors of the product, or an estimate was based on the indexes for similar products purchased by other laboratories.

In the private laboratories, the problem of pricing nonhomogeneous products arose frequently, and has not yet been solved satisfactorily, though several approaches have been proposed. A development contract calls for the delivery of hardware and, especially if the device being developed

is a weapon system of considerable magnitude, the major components, or subassemblies, are regularly farmed out for development to subcontractors. Such a component may be produced only once; or it may undergo such extensive redesign that any attempt to adjust its price is futile. In other cases, the prices quoted in the second and subsequent years (which do not include the subcontractor's learning and tooling costs) are so disproportionate to the initial price that they exaggerate the price decline.

One reason for these sharp price declines is the difference between research and development for the industrial or consumer market, on the one hand, and R&D for military use, on the other. The developer for the consumer market produces and sells the product he develops. With the aid of market research, or on the basis of past experience, he can predict to some extent the potential market for his product. Development costs can then be included in the unit price on a prorated basis and liquidated over a relatively long time span. R&D performed on military products, however, is an end in itself, and the contractor must recoup the costs during the life of the R&D contract. If the item later goes into production, a new contract will be drawn which has no bearing on the R&D effort.

In this pilot study, the best response with respect to pricing these unique, nonhomogeneous products was obtained from a prime contractor who in turn obtained estimated price movements based on the subcontractor's year-to-year cost experience for labor, overhead, and materials. In other cases, specification differences were priced directly by the subcontractor. The latter method may provide a better longrun solution.

From the viewpoint of precise pricing, the best approach would be to compare specification changes and to price the changes. This is the method used for the Bureau's Wholesale Price

8 Sometimes this difficulty could be rectified by pricing items which were important in the judgment of the laboratories. Analysis of the reports of private laboratories indicated that approximately 25 percent of the prime contract dollar was passed on to subcontractors and that costs incurred from subcontractors accounted for 60 to 67 percent of the prime con tractors' FY 1961 materials expenditures. This points up the need to include subcontractors in the compilation of an R&D index.

10 This approach has been investigated and found to be feasible. The primary objection is that it does not take into consideration productivity gains at the subcontractor level, thus injecting pos sible upward bias in the total index.

Index and, when applied, eliminates price variations resulting from quality changes. In an R&D index, this approach would correctly reflect the productivity improvements at the subcontractor level.

Another solution would be to treat subcontractor-furnished material separately. This idea includes expenditure and pricing surveys for the subcontractors as well as the primes. Those subcontractors found to be engaged in R&D work could be included in the survey sample, or the indexes obtained are used to alter the applicable values listed in the prime contractor's expenditure report.10

Preliminary Results

While appearing reasonable in trend, the indexes should not be taken as precise measures inasmuch as a number of statistical problems remain unsolved. (See table.) However, some observations can be made. The difference between the private and Army laboratory indexes for materials in 1962 was due largely to sharp increases in prices paid by certain Army laboratories located at considerable distances from supply centers. The effect of the pay increase received by Federal employees in late 1962 showed up in the Army direct abor index in 1963 (5.8 percent, compared with the 5.6 percent reported for all Classification Act employees).

This experimental study reveals that it is feasible to construct a reliable index of research and development cost prices given sufficient resources and cooperation of reporters. Input price indexes of this general type could be useful as a measure of the purchasing power of R&D funds and, ultimately, as a price deflator to be used in obtaining a constant-dollar series for measuring volume changes in R&D (assuming no change in productivity). Indexes of price trends for various types of R&D effort, classes of personnel, and other cost elements could also be constructed. Though the indexes will not measure changes in total R&D costs or changes in the price of a unit of R&D output, they will be useful in the construction of additional indexes for such measurements.

In order to improve the measures, the sample of private contractors should not be limited to the largest prime contractor of an intramural laboratory, but rather additional primes should be selected to represent diverse activities. Whether the type and volume of data will support a probability sample is by no means certain. Subcontractors who supplement the activities of prime contractors should be included, as this would provide a more complete picture of the relationship between labor and material costs, and, would provide a means for dealing with at least part of the problem of pricing nonhomogeneous products.

796-484 0-66

Significant Decisions in Labor Cases*

Labor Relations

Secondary Boycott. Applying its traditional rules on picketing in common situs situations, based on consideration of timing and proximity of picketing to the struck employer's site of operations, the National Labor Relations Board found 1 a union to have unlawfully picketed construction site gates assigned for the exclusive use of neutral employees. In a split decision, the Board's majority declined to apply a test developed by the U.S. Supreme Court, which permits separate gate appeals to neutrals if their work is "necessary to the normal operations" of the primary employer.

The main contractor for the expansion of a filtration plant petitioned the Board to stop the union of its employees from picketing construction site gates reserved exclusively for the employees of neutral subcontractors. It contended that the picketing was unlawful because it was designed to, and did, enmesh the neutral employees in the primary dispute, in violation of the Labor Management Relations Act.

The union responded that its picketing of the reserved gate was in furtherance of its demands upon the primary employer and was protected by the "primary strike and picketing" proviso of the act. It further contended that since the work of the subcontractors was on the same jobsite and was related to employer's normal construction operations, its picketing was lawful primary activity under the test developed by the U.S. Supreme Court in the General Electric case.2

The Board held that the General Electric principle is inapplicable where the primary contractor is but one of several employers on premises owned or operated by another party-that is, in a common situs situation such as that involved in the present case.

Instead, the Board held that the standard established by it in the Moore Dry Dock Co.3 decision was applicable in determining whether the site of the picketing was reasonably close to the operations of the primary employer, and whether a union picketing a common situs had taken reasonable precaution to prevent enmeshment of neutral employees and contractors in the primary dispute. In the instant case, the Board held, the union did not meet the test and was in violation of the act.

Dissenting Members Fanning and Jenkins saw no reason for the majority's declining to apply the General Electric test to the construction industry. They held that certainly Congress did not intend a more lenient treatment to be given neutrals at an industrial site than neutrals at a building construction site. Especially is this true where the work of the neutrals is related to the normal operations of the primary employer, as in the instant case, the dissenters said.

The minority held that the General Electric test was applicable to both industrial and construction disputes and should have been used in this case.

*Prepared in the U.S. Department of Labor, Office of the Solicitor. The cases covered in this article represent a selection of the significant decisions believed to be of special interest. No attempt has been made to reflect all recent judicial and administrative developments in the field of labor law or to indicate the effect of particular decisions in jurisdictions in which contrary results may be reached based upon local statutory provisions, the existence of local precedents, or a different approach by the courts to the issue presented.

1 Building and Construction Trades Council and Markwell and Hartz, Inc., 155 NLRB No. 42 (Oct. 25, 1965).

2 Electrical Workers Local 761 v. Labor Board, 366 U.S. 667; see also Monthly Labor Review, August 1961, p. 879. In this decision, the Supreme Court ruled that picketing at a gate used exclusively by neutral employees working on the primary employer's premises is lawful unless there is "a separate gate, marked and set apart from other gates; the work done by the men who use the gate [is] unrelated to the normal operations of the employer; and the work [is] of a kind that would not, if done when the employer were engaged in its regular operations, necessitate curtailing those operations."

392 NLRB 547. There the Board set the following test of lawful picketing of premises occupied by secondary employees: (a) the picketing is strictly limited to times when the situs of the dispute is located on the secondary employer's premises; (b) at the time of the picketing the primary employer is engaged in its normal business at the situs; (c) the picketing is limited to places reasonably close to the location of the situs; and (d) the picketing discloses clearly that the dispute is with the primary employer.

4 Fabri-Tek, Inc. v. NLRB (C.A. 8, Nov. 5, 1965).