The United States Department of Labor (the DOL) and defendant-intervenors, Ce-drick Turner, et al„ appeal from the District Court’s order of August 22, 1984, 597 F.Supp. 45, holding (1) that the regulation promulgated by the DOL at 20 C.F.R. § 655.207(b),1 as amended by the DOL on August 31, 1983, and found at 48 Fed.Reg. 40.168 (September 2, 1982) and known as the Adverse Effect Wage Rate (AEWR), to be arbitrary and capricious and thereby invalid, and (2) directing the Secretary of Labor to promulgate an AEWR for the year 1983 pursuant to the rule-making provisions of the Administrative Procedure Act (APA, 5 U.S.C. § 706) and consistent with the Court’s opinion filed therein. We conclude that the District Court erred in denying the DOL’s motion for summary judgment, and accordingly reverse.
I.
The issues on appeal arise in two eonsol-idated suits brought by certain associations and growers of various agricultural prod-acts in Virginia, Maryland, and New York, seeking to invalidate and have set aside the 1983 AEWR published by the DOL on September 2, 1983, and the methodology adopted by the DOL on that date for making annual adjustments to the AEWR for subsequent years.
In seeking injunctive and declaratory relief, plaintiffs specifically attack the adoption by the DOL of the methodology for adjusting the AEWR by utilizing data procured from what was designated as the ES-202 program. This program is a cooperative activity of the Bureau of Labor Statistics (BLS) and state employment security agencies. As part of their unemPayment insurance programs, these state agencies receive reports from employers covered by unemployment insurance indicating, inter alia, the number of workers on the payroll, total wages, and taxable wages. The agencies, in turn, report quarterly to the BLS on the number of covered establishments, employment during the mid-week of each month, and total wages paid during the quarter. The wages are reported to the BLS using Standard Industrial Classification (SIC) codes, including categories of agricultural crop producers, Although the precise coverage of agricultural labor afforded by the data varies from state to state, it encompasses, at a minimum, employees of agricultural firms employing at least ten workers in twenty weeks and having a $20,000 quarterly payroll.
*92In response to the allegations of the complaints, the DOL moved for summary judgment on the ground that the full administrative record before the District Court for review satisfied the requirement that the agency had considered all relevant factors in discharging its rule-making responsibilities and had complied with the provisions of the Administrative Procedure Act (APA), 5 U.S.C. § 706.
Subsequent to the District Court's denial of the DOL’s renewed motion for summary judgment on May 17, 1984, a two-day trial was held by the Court, at which plaintiffs presented two witnesses, both of whom were consultants employed by them and whose testimony challenged the regulation and suggested alternatives. At the conclusion of the trial, the District Court issued the opinion and order, which is the subject of this appeal. The trial court concluded that “a” “[tjrial was necessary in order to determine whether the DOL considered all relevant factors and in order to discharge properly [the court’s] duty to engage in ‘substantial inquiry’ as to whether the process employed by the DOL to reach its decision took into consideration all relevant factors.”
II.
On appeal the DOL argues that by conducting a de novo hearing to determine the issue presented, the District Court exceeded its limited function as a reviewing Court. The agency contends that its action was a reasonable exercise of its rule-making authority. Further the DOL urges that its position is supported by the administrative record and even the specific language of the District Court’s Memorandum Opinion.
III.
The trial court’s conclusions must be tested in light of the record before it and the large body of case law touching every facet of review of administrative action by the courts.
The Supreme Court has emphasized that “the focal point for judicial review should be the administrative record already in existence, not some new record made initially in the reviewing court.” Camp v. Pitts, 411 U.S. 138, 143, 93 S.Ct. 1241, 1244, 36 L.Ed.2d 106 (1973). This court has observed that “de novo review is appropriate only in special circumstances where agency fact finding procedures are inadequate in an adjudicatory proceeding, or where issues not before an agency are raised in a proceeding to enforce nonadjudicatory agency action.” United States v. Holcomb, 651 F.2d 231, 236 (4th Cir.1981), citing Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415, 91 S.Ct. 814, 824, 28 L.Ed.2d 136 (1971). Neither of these two narrow exceptions to the general proscription of de novo review was present in this case. We believe that in holding a trial to determine whether the agency had considered all relevant factors, the District Court failed to observe the Supreme Court’s declaration in Camp, supra, that if “there was such failure to explain administrative action as to frustrate effective judicial review, the remedy was not to hold a de novo hearing but to obtain from the agency, either through affidavits or testimony, such additional explanation of the reasons for the agency decision as may prove necessary.” Camp, supra, at 142-143, 93 S.Ct. at 1244. Instead of seeking agency supplementation of the administrative record where required or suggested, the court allowed the same to be supplemented with testimony from plaintiffs’ consultants, challenging the wisdom of the regulation and suggesting alternative approaches that might have been developed but were not before the agency when it made its decision. Furthermore, the court repeatedly relied upon this testimony as a basis for its conclusions in its memorandum opinion.
As was so aptly observed by Judge Wilkinson in Berry v. Ciba-Geigy Corporation, 761 F.2d 1003 (1985), “the standard exists to ensure that administrative responsibility rests with those whose experience is daily and continual, not with judges whose exposure is episodic and oc*93casional.” Berry, supra, at 1006. Persuasive on the issues before us also are the cases of Asarco, Inc. v. United States E.P.A., 616 F.2d 1153, 1161 (9th Cir.1980) and Proietti v. Levi, 530 F.2d 836, 838 (9th Cir.1976), in which the Ninth Circuit held that district courts are generally accorded no deference in their review of agency actions where review is limited to the administrative record. That is the precise situation here. The trial court was obligated to look to the administrative record and its view of the record is not entitled to deference. Its action is to be weighed in light of the appropriate standards dictated by the Administrative Procedures Act and the applicable case law.
IV.
Since the scope of review of the DOL’s rule-making action is governed by the standards set forth in 5 U.S.C. § 706(2)(A), the courts must uphold agency action unless it was “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” Id. Under this standard a reviewing court must decide if the agency’s decision “ ‘was based on a consideration of the relevant factors and whether there has been á clear error of judgment____ Although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The Court is not empowered to substitute its judgment for that of the agency.’ ” Shoreham Cooperative Apple Producers Ass’n., Inc. v. Donovan, 764 F.2d 135 (2nd Cir.1985) (quoting Citizens to Preserve Overton Park, Inc. v. Volpe, supra 401 U.S. at 416, 91 S.Ct. at 823-24.) The agency is obligated to examine the available evidence and to articulate a “rational connection” between that evidence and its exercise of discretion. Motor Vehicle Manufacturers Ass’n. v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866-67, 77 L.Ed.2d 443 (1983) (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S.Ct. 239, 245, 9 L.Ed.2d 207 (1962)). Far from being deficient, the administrative record in this case amply explains the DOL’s reasons for abandoning the prior methodology and for deciding to use ES-202 data. The DOL offered cogent reasons why it could no longer rely on farm wage data collected by the United States Department of Agriculture. Furthermore, the alleged flaws in the ES-202 base set forth by the appellees do not demonstrate that the DOL’s action was arbitrary and capricious.
Shoreham Cooperative Apple Producers Ass’n., Inc. v. Donovan, hereinabove mentioned, is of particular importance. In this case the Second Circuit granted an interlocutory appeal to consider the denial of the DOL Secretary’s motion for summary judgment in a suit attacking the validity of the same regulation before us. The Second Circuit reversed with instructions to enter summary judgment for the Secretary, ruling that the record contained “no basis for holding that the DOL failed to consider available evidence or otherwise made its choice arbitrarily.”
Based on the administrative record before the District Court, which contains clear and reasonably detailed explanations of the agency’s reasons for the choice made, and in light of the Second Circuit’s holding in Shoreham, supra, we find no basis for holding that the DOL failed to consider available evidence or otherwise made a choice that was arbitrary and capricious. We hold, therefore, that the District Court erred in denying the DOL’s motion for summary judgment. Accordingly, this case is remanded to the District Court with instructions to enter summary judgment in favor of the DOL.
REVERSED AND REMANDED.
. 20 C.F.R. § 655.207(b) provides in pertinent part:
[T]he adverse effect rate for each year shall be computed by adjusting the prior year’s adverse effect rate by the percentage change (from the second year previous to the year previous) in the ES-202 report’s aggregate average weekly wage rates for the appropriate group of agricultural workers. The appropriate group of workers shall be those U.S. agricultural workers employed by establishments in Standard Industrial Classification (SIC) Code Nos. 013, 016, 017, 019, 071, and 072 within that State (except that for purposes of wage movement, but not actual adverse effect rates, New York, Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont shall be considered as one State, and Maryland, Virginia, and West Virginia shall be considered as one State.) The Administrator shall publish, in each calendar year, on a date he shall determine, adverse effect rates calculated pursuant to this paragraph (b) as a notice in the Federal Register.