MEMORANDUM
HAROLD H. GREENE, District Judge.In this action, several non-profit organizations which had been designated as grantees under section 402 of the Job Training Partnership Act (JTPA), 29 U.S.C. § 1672,1 challenge the allocation of funds for programs under the Act. The statute authorizes the Department of Labor (DOL) to award grants to public agencies and non-profit organizations to support programs and activities for migrant and seasonal farmworkers.2 It appears that the allocation to California and several other states for fiscal year 1984 is substantially below their fiscal year 1983 allocation, while that of several other states is being proportionately increased.3 Plaintiffs challenge the change in allocation on various grounds.
Prior to fiscal year 1984, DOL used Social Security Administration (SSA) data as the basis upon which to distribute funds for farmworker job training programs. However, section 162(a) of the JTPA provides that:
All allotments and allocations under this Act shall be based on the latest available *697data and estimates satisfactory to the Secretary. All data relating to economically disadvantaged and low-income persons shall be based on 1980 Census or later data.
In accordance with this congressional mandate, DOL shifted over from SSA data to 1980 Census data and that shift, in addition to certain other base data decisions discussed below, resulted in the redistribution of funds to which plaintiffs object.
Plaintiffs argue that the methodology used by DOL was improper on four different bases.4 The Court finds no arbitrary action with respect to any of the choices made by DOL.
First. Plaintiffs claim that DOL erred in requesting the Bureau of the Census to furnish it numbers of individuals from ten occupational codes, and they suggest that the Department should have instead used industrial codes similar to those in the SSA 1979-83 data base.5
Since the Census Bureau does not maintain a definition of “farmworker” or data for “farmworker population,” DOL necessarily had to make a decision on the most appropriate method for satisfying the congressional mandate. The DOL employee selected for developing the necessary figures, who had ten years of experience in farmworker programs and familiarity with various data bases for allocation purposes, conferred with Census and with other DOL employees, and he ultimately selected ten Census codes for farm and related agricultural occupations which in his view best captured all persons who were attempting to make a living in agriculture on a seasonal basis. This method of proceeding was far from arbitrary; indeed, it appears to be considerably more reasonable and accurate than plaintiffs’ proposed use of the industrial codes in the SSA data base which captured many non-farmworkers such as clerical, administrative, and technical employees of agricultural firms, and therefore vastly overstated the farmworker population.6
Second. Plaintiffs argue next that DOL improperly included in its computation self-employed farm operators notwithstanding the fact that the regulatory definition of “farmwork” is “work performed for wages” in agricultural production or services. 20 C.F.R. § 633.104. DOL has explained that self-employed persons were included because this category comprehends many tenant farmers and sharecroppers who are among the poorest individuals attempting to make a living in farmwork and who regularly augment their income by working for wages for other farmers. In response to plaintiffs’ claim that the inclusion of self-employed persons would count individuals who might be operating large agribusiness enterprises, DOL points out, with considerable justification, that this could not occur because of the additional poverty-level requirement in the statute and the regulations. On the basis cited, the inclusion of self-employed persons was not arbitrary.
Third. Plaintiffs complain that DOL should have requested Census to identify those individuals who were below 70 percent of the lower level standard income level (LLSIL) or on public assistance, in view of the fact that section 633.107(a) of the regulations includes these categories in defining eligibility for participation in the program.
DOL points out in response that (1) the Census reported that it would take a considerable amount of time to provide the LLSIL information, and that the overall data base for that reason might not have been ready in time for the allocation of FY *6981984 had it awaited these figures, (2) DOL had never in prior years attempted to identify those who were below 70 percent of the LLSIL or on public assistance but had then, as now, applied instead, without challenge from any grantee, a poverty level factor of $3,000; and (3) perfect symmetry between eligibility criteria and the allocation data base is not required under the statute or the regulation.7 These arguments, too, are persuasive in view of the difficult choices which DOL had to make.
Fourth. Plaintiffs argue that DOL should have requested the Census to supply it with information on dependents who are eligible individuals. DOL notes in response that there is no basis for assuming that dependents would be more concentrated in particular geographic areas.8 Moreover, in this regard, too, the Department had never in prior years included dependents and it had never been challenged with respect to that decision by any grantee.
It is unquestionably possible to find fault with the methodology used by the Department of Labor. In particular, refinements and adjustments might well have been made with respect to those aspects of the formula which do not conform the allocation formula to the eligibility criteria as precisely as possible, whether or not they had been challenged in previous years. However, these are relatively minor flaws in an area where DOL had to make complex decisions in a relatively short period of time. It appears, moreover, that plaintiffs’ quarrel is primarily with the substitution of the 1980 Census data for the 1977 SSA data, and that choice must be laid at the doorstep of the Congress.
Certainly, the method used in the past to allocate funds among the states was far less suited to locate the target population than is the present formula, and the Congress required use of the more recent Census data precisely for that reason. Plaintiffs’ plan — to have the Court declare illegal the present system used by DOL and to require a return to that in force prior to FY 1984 until a more perfect allocation formula can be devised — is far less consistent with the will of Congress than the method actually used by DOL.
The Court’s role is limited. At worst from plaintiffs' point of view, the Secretary’s choice of data base is entirely non-reviewable;9 at best, that choice may be overturned only if the Court finds it to be arbitrary or capricious.10 In making its decision in that regard, the Court is, of course, required to defer to the agency’s expertise.11 On the basis of the explanation provided supra, the Court concludes that the Secretary carefully sought to arrive at an allocation formula based on factual data'reasonably available to him, and that his actions cannot be characterized as arbitrary or capricious.12 The action will be dismissed.
. This statute replaced the Comprehensive Employment and Training Act (CETA), 29 U.S.C. §§ 801-992.
. The purpose of the statute, like that of its CETA predecessor, is to assist migrant and seasonal farmworkers and their families to improve their living conditions and to develop skills necessary for a productive and self-sufficient life.
. California’s share was reduced from 21.69 percent to 8.80 percent, Arizona’s from 2.21 percent to 1.07 percent, and Massachusetts from .64 percent to .33 percent. By contrast, South Dakota’s share increased from .47 percent to 2.48 percent, Kentucky’s from 1.16 percent to 3.26 percent, and Iowa’s from 1.86 percent to 4.12 percent.
. Plaintiffs also claim that DOL ignored its own regulations which require that allocations be allotted for farmworkers programs in individual States "in an equitable manner.” 20 C.F.R. § 633.105(b).
. 20 C.F.R. § 633.104.
. It should be emphasized that the total amount available for farmworker training programs is not at issue here. However DOL may have made its computations, that amount was fixed by the congressional appropriation. Only the distribution between and among the states is being challenged in this lawsuit.
. DOL recognizes that there must be a reasonable relationship between the data base for allocation purposes and the individual eligibility requirements. Reply Memorandum at 13. It is not necessary in this case to explore in what circumstances that relationship would be so far distant as to be no longer reasonable. In view of the many imponderable factors, perfect symmetry was not achievable, and the "reasonable relationship" standard was therefore appropriate.
. Program operators are not permitted to spend more than 15 percent of their grant funds for non-training related supportive services, such as services for dependents. 20 C.F.R. § 633.304.
. 29 U.S.C. § 1572(a) places no guidelines on the Secretary's choice of an allocation formula other than the requirement that data relating to low income individuals shall be based on the 1980 Census or later available data. Indeed, the statute provides that the data chosen shall be “satisfactory to the Secretary." See Natural Resources Defense Council v. SEC, 606 F.2d 1031, 1043 (D.C.Cir.1979).
. Section 10(e)(B)(l) of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A).
. See National Association of Metal Finishers v. EPA, 719 F.2d 624, 657 (3rd Cir.1983); State of Connecticut v. EPA, 696 F.2d 147, 159-60 (2d Cir.1982).
. Plaintiffs’ procedural claims are likewise without merit. DOL stated in its announcement that its new state allocations would be based on 1980 Census data on agricultural occupations. 48 Fed.Reg. 36688. Thereafter, the grantees, *699including plaintiffs, attended a conference during which the change in the data base for state allocation purposes, as well as DOL’s methodology and rationale, were extensively discussed. See Forester v. CPSC, 559 F.2d 774, 787 (D.C.Cir.1977); Arlington Oil Mills, Inc. v. Knebel, 543 F.2d 1092, 1099-1100 (5th Cir.1976). Subsequent comments relating primarily to the substitution of the 1980 Census data for the 1977 SSA data, and DOL fully responded to that issue. 48 Fed.Reg. 54914. DOL was not required to respond to the details of every comment. Action on Smoking and Health v. CAB, 699 F.2d 1209, 1216 (D.C.Cir.1983).