In February, 1996, the Wage and Hour Division initiated Survey No. 96-PA-003, a wage survey of residential construction projects in Allegheny County, Pennsylvania (including Pittsburgh). All interested parties, including Plumbers Local 27, were advised of the survey and invited to submit data. The cut-off date for submitting data was April 5, 1996. Administrative Record (AR) Tab C.

AR Tab C.

   By letter to the Wage and Hour Administrator dated December 16, 1996, Local 27 formally requested review and reconsideration of the wage determination for plumbers performing work on residential projects in Allegheny County, PA, as provided by 29 C.F.R. §1.8 (1997). AR Tab B. On April 24, 1997, the National Office Program Administrator affirmed the earlier wage determination.

   The Davis-Bacon Act charges the Secretary with responsibility for determining locally prevailing wage rates for various classes of workers employed on construction projects. The Secretary's regulations for determining prevailing wage rates are found at 29 C.F.R. Part 1. In addition to the regulations, the Wage and Hour Division has developed two guidance documents that are used internally in the wage determination process: (a) the Davis-Bacon Construction Wage Determinations Manual of Operations (April, 1986)(Manual of Operations); and (b) an internal training manual.

   The Division has an on-going program of compiling wage data for issuing updated wage rates. Information on wage rates for the various job classifications is solicited periodically from contractors, contractors associations, unions, governmental sources, and others in a given locality. 29 C.F.R. §1.3(a). This process primarily involves developing data identifying individual jobsites in a locality, the names of the contractors working at the site, project cost, dates of project start and completion, as well as the number of employees working at the site, their classification and wage/fringe benefit rates. 29 C.F.R. §1.3(b)(1). Typically, these surveys (and the resulting wage determinations) are divided into four broad categories of construction: residential, building, highway and heavy. Manual of Operations at 23-29, citing All Agency Memoranda 130 and 131.

   The Administrator has developed internal guidelines for determining whether enough reliable data has been submitted during the survey process to support the issuance of a wage determination. First, the Administrator calculates the useable response rate, defined as the number of responses received that provide useable survey data divided by the total number of contacts made with general contractors and subcontractors. AR Tab C (Manual of Operations) at 62. The Manual of Operations continues with this guidance:

In most instances, when survey response rate equals at least 25%, data for a substantial number of workers in each of the major classes will be received. However, for surveys conducted in rural counties where construction activity is sparse or for highly specialized classes, the number of employees for whom data are provided may be limited. In such cases, a wage rate for an individual class is to be recommended only when information on at least six workers is received from three or more contractors, none of which accounts for 60% or more of total reported employment [Footnote omitted]. However, if the overall survey usable response rate is 50% or more, data on only three workers, from two contractors may be utilized. Exceptions to these standards may be allowed, but must be justified in writing.

29 C.F.R. §1.3(d).

   The Administrator asserts that because the Allegheny County residential survey produced data on at least 3 workers employed by 2 contractors on privately-funded residential work, the private sector data was sufficient per se under the Division's internal policy guidelines, and the Administrator therefore was free to ignore entirely the data on more than 505 plumbers on federally-funded residential work. We respectfully disagree.

47 Fed. Reg. at 23645 (emphasis added). Although the Secretary did not specifically express a rationale for this limitation on the use of data from Davis-Bacon projects when issuing the regulation, the Secretary noted in the final rulemaking that the proponents of this change were concerned that the inclusion of wage data from federally-funded building- and residential-type projects tended to skew the Department's survey results upward when compared with rates typically paid on privately-funded projects in the locality. Id.

   Nowhere in the regulation is the term "sufficient data" defined, suggesting that the Administrator has discretion in making the determination whether the available wage data from private sector projects is insufficient to issue a wage rate without also incorporating wage data from projects subject to Davis-Bacon wage standards. However, it is the Board's responsibility to insure that the exercise of this discretion is in keeping with purposes of the Act and the regulations.

   Definitionally, the word "sufficient" suggests that a criterion is to be evaluated in a specific context, i.e., in relationship to some purpose. Black's Law Dictionary defines "sufficient" as "[a]dequate, enough, as much as may be necessary, equal or fit for end proposed, and that which may be necessary to accomplish an object. Of such quality, number, force, or value as to serve a need or purpose." Black's Law Dictionary 1443 (6th ed. 1990). An amount or quantum that may be "sufficient" for one purpose plainly might be insufficient in another context. The underlying "object" or "purpose" of the wage determination process is the Secretary's statutory mandate to determine the locally prevailing rate of pay for laborers and mechanics engaged on construction projects, 40 U.S.C. §276(a), as guided by the various considerations found within the Secretary's implementing regulations at 29 C.F.R. Part 1.

   When deciding whether a wage survey has developed sufficient data to justify the publication of a prevailing wage rate for any particular job classification, the Administrator has adopted a threshold standard for sufficiency. The Administrator will publish a wage determination for a particular job classification so long as usable responses have been received on at least three workers employed by at least two employers (the "3 worker/2 contractor rule"). According to the Administrator, this minimal standard is based upon guidelines used by the Bureau of Labor Statistics. AR Tab C (Manual of Operations) at 62. In the hearing before the Board, counsel for the Administrator represented that this relatively low standard has been adopted in part because it often is difficult to develop large amounts of wage data in a given locality as part of the survey process, particularly in connection with residential construction surveys.

   The Wage and Hour Division's general use of this fixed 3 worker/2 contractor standard when deciding whether there is sufficient wage data to publish a wage determination rate is not at issue in this proceeding. However, the Administrator would import this same fixed standard to the determination whether the available private sector data is "sufficient" under 29 C.F.R. §1.3(d), i.e., whether data from federally-funded projects should be excluded from consideration when calculating a wage determination because there is adequate data from privately-funded projects. In light of the facts before us in this case (in which data from six plumbers on privately-funded work has resulted in wage data on 505 other plumbers being excluded entirely from consideration), it is our view that the Administrator's wholesale reliance on the fixed 3 worker/2 contractor formula is misplaced because large amounts of relevant wage data on federally-funded projects are excluded based solely on a de minimis showing of data from privately-funded jobs.

   This Board's predecessor, the Wage Appeals Board, addressed a similar issue in Southeast Idaho Building and Construction Trades Council, AFL-CIO, WAB Case No. 86-22 (Feb. 4, 1987), a case in which Wage and Hour had concluded that there was insufficient construction data in two rural counties from which to issue a wage determination under 29 C.F.R. §1.7(b). Section 1.7(b) provides that if there is insufficient wage data available in a rural county to issue a wage determination, the Administrator is authorized to enlarge the geographic scope of the survey area to include adjacent rural counties. In Southeast Idaho, the Administrator had decided to consider only wage data from privately-funded projects in the locality when determining the sufficiency of data, and to exclude data from federally-funded projects as part of the survey analysis. After concluding that the data only from privately-funded projects was insufficient to support the issuance of a wage determination, the Division expanded the geographic scope of the survey area to include other surrounding rural counties in the state. In a divided decision, the Wage Appeals Board rejected this approach, finding that:

in the first instance all wage data should be collected whether on private or public construction projects. It is not until after compiling wage rate data can it be determined whether to exclude Federal and federally-assisted projects subject to the Davis-Bacon prevailing wage requirements.

Southeast Idaho, slip op. at 4. In short, the Board held in Southeast Idaho that it is only after the Administrator reviews all the relevant data before the Division (i.e., data that is within the scope of the survey in terms of time, type

of construction, geographic area, etc.), whether from federally- or privately-funded projects, that the Administrator can make a determination concerning the sufficiency of wage data under Section 1.7(b). In Southeast Idaho, the Board remanded to the Wage and Hour Administrator to issue separate prevailing wage determinations for the two counties in question, based on data from both private and federally-funded projects that were included in the survey.