McGOWAN, Senior Circuit Judge:

The question presented in this case is whether the Federal Mine Safety and Health Review Commission properly reversed an administrative law judge's (ALJ's) finding that Phelps Dodge Corporation illegally discriminated against an employee active in promoting mine safety when it twice disciplined him for alleged on-the-job infractions. Unlike the Administrative Procedure Act, 5 U.S.C. § 557(b) (1976), the Commission's generic statute limits the agency's review of an ALJ's findings of fact to an inquiry into whether they are supported by substantial evidence. 30 U.S.C. § 823(d)(2)(A) (Supp. V 1981). Because the ALJ's findings in this case were firmly supported by the record, we hold that the Commission misapplied its statutory standard of review. Accordingly, we reverse the Commission's decision.

I

Before 1979 Johnny Chacon had worked for approximately ten years as a locomotive operator at a Phelps Dodge open-pit copper mine. During that time he received various forms of discipline roughly once per year for job-related infractions. See Joint Appendix (J.A.) 322-26. Shortly before January 1979, when Chacon became vice-chairman of the union local, he began to press the union's safety complaints more vigorously than the union had ever done. During a thirty-five-day period soon after he began his safety campaign, Chacon received three disciplinary warnings and a three-day suspension without pay. This case concerns two of the four disciplinary actions from this period. On February 5, 1979, Chacon derailed his train, as operators frequently did on the portable track at the mine. He received a written warning the next day allegedly for having travelled at excessive speed while a slow order was posted. Six days later Chacon again derailed a train. He was cited for the same infraction and this time suspended for three days without pay.

The Secretary of Labor filed a complaint on Chacon's behalf alleging that these disciplinary actions amounted to discrimination against Chacon's safety activism and were therefore illegal under section 105(c)(1) of the Federal Mine Safety and Health Act of 1977, 30 U.S.C. § 815(c)(1) (Supp. V 1981) ("Mine Act"). After a hearing, an ALJ found in the Secretary's favor. Secretary of Labor ex rel. Chacon v. Phelps Dodge Corp., 2 F.M.S.H.R.C. 1271 (ALJ 1980) ( "ALJ decision"), rev'd, 3 F.M.S.H.R.C. 2508 (1981). It ordered Phelps Dodge to expunge the disciplinary actions from Chacon's personnel file and ordered the company to pay lost wages and a civil penalty of $2500. The ALJ found the penalty justified because the offense was a "very serious violation" that would have a "dampening effect" on safety reporting at the mine. Id. at 1288.

The Commission granted Phelps Dodge's petition for discretionary review of the ALJ's decision. Although it ultimately reversed the ALJ, Secretary of Labor ex rel. Chacon v. Phelps Dodge Corp., 3 F.M.S.H. R.C. 2508 (1981) ( "Commission decision"), the Commission began by endorsing his conclusion that the Secretary had established a prima facie case of discrimination. Quoting a portion of its holding in Secretary of Labor ex rel. Pasula v. Consolidation Coal Co., 2 F.M.S.H.R.C. 2786, 2799 (1980), rev'd on other grounds sub nom. Consolidation Coal Co. v. Marshall, 663 F.2d 1211 (3d Cir. 1981), the Commission explained that a prima facie case is established "`if a preponderance of the evidence proves (1) that [the miner] engaged in a protected activity, and (2) that the adverse action was motivated in any part by the protected activity.'" Commission decision, 3 F.M.S.H.R.C. at 2509. It was not seriously disputed that the Secretary had established the first prong. As for the second part of the test, the Commission agreed with the ALJ's conclusion that the discipline was motivated at least in part by a desire to retaliate for Chacon's safety activities. It found persuasive the Secretary's evidence that the company knew of Chacon's safety activities, that it was hostile toward them, and that the discipline came almost immediately after the safety activities began. Id. at 2510-11.

Central to the Commission's holding in the company's favor, however, was Pasula's statement that a company may defend against a prima facie case of discrimination "`by proving by a preponderance of all the evidence that, although part of [its] motive was unlawful, (1) [it] was also motivated by the miner's unprotected activities, and (2) that [it] would have taken adverse action again[s]t the miner in any event for the unprotected activities alone.'" Id. at 2514 (quoting Pasula, 2 F.M.S.H.R.C. at 2799-800 (original underscores "would")). The ALJ had offered several reasons to support his finding that the company would not have disciplined Chacon in the same way had he not been a safety activist. First, the ALJ noted that discipline for derailments accompanied by speeding was exceedingly rare. In 1977, 1082 trains derailed at the mine; in 1978, 1164 did. Nonetheless, in 1977 the company issued no warnings for excessive speed and in 1978 issued only four. From 1976 to 1979, the company issued only one suspension for speeding in addition to Chacon's. ALJ decision, 2 F.M.S.H.R.C. at 1274, 1276. Moreover, the ALJ found evidence consistent with the view that when the company did resort to punishment, the offenses were more serious than Chacon's. Estimates of Chacon's speed at the time of his two derailments ranged between ten and fifteen miles per hour. See infra pp. 89-90. In contrast, although only one of the four previous warnings listed the operator's speed, in that case the operator was traveling at twenty miles per hour. Two of three warnings issued after Chacon's discipline listed the operator's speed: fifteen and twenty miles per hour. ALJ decision, 2 F.M.S.H.R.C. at 1276.

Second, the ALJ thought that disagreement between employees and supervisors at Phelps Dodge over the meaning of a slow order was evidence that the company did not consistently punish operators in Chacon's situation. In the absence of a slow order, the maximum speed was fifteen miles per hour. Chacon and another employee testified that the company would often specify a reduced maximum speed when it posted a slow order. When it did not do so, however, they understood that no specific maximum speed was implied; rather, the appropriate speed varied according to track conditions and was left to the operator's good judgment. See J.A. 60-61, 105-06. Supervisors asserted that a ten-miles-per-hour maximum implicitly inhered in any slow order, though they disagreed even among themselves on the slow order's precise meaning. The ALJ concluded that the meaning of a slow order was so vague that the company had "complete latitude" in later proceedings to attach any meaning that then suited its purpose. ALJ decision, 2 F.M.S.H.R.C. at 1283. This vagueness led the ALJ to find unbelievable that the company consistently enforced a specific speed limit on slow orders that did not specify a maximum speed.

Finally, even if ten miles per hour were the implicit maximum, the ALJ found that Chacon had not exceeded that speed. True, the tape from the train's speedometer indicated speeds of sixteen and fifteen miles per hour at the time of the two derailments. But the ALJ thought it well-established that the speedometers were highly unreliable: testimony showed that the speedometer's needle bounced erratically between five and fifteen miles per hour. J.A. 55. Moreover, the ALJ found that the company could not have inferred Chacon's speed from the damage he caused. Company testimony that one could so calculate a train's speed was not corroborated by any facts the company offered, and the ALJ found evidence to the contrary in the fact that damage varied according to factors unrelated to a train's speed, such as track conditions. ALJ decision, 2 F.M.S.H.R.C. at 1278. Finally, Chacon and his supervisor agreed immediately after the February 12 derailment that he had been traveling between ten and twelve miles per hour when the incident occurred, id., and the ALJ concluded on the basis of Chacon's own testimony that the lower figure was the best estimate, id. at 1277.

The Commission disagreed and thus reversed the ALJ. It thought "Phelps Dodge established . . . that it would have disciplined Chacon in any event for the unprotected activities alone." Commission decision, 3 F.M.S.H.R.C. at 2517. First, it questioned the significance of the evidence showing that Phelps Dodge issued warnings or suspensions for speeding only infrequently. It was possible, the Commission said, that the company warned only four operators for speeding in 1977 and 1978 because only four of the more than 2000 trains that derailed during those two years both caused damage and were speeding. The Secretary did not directly identify any speeders who had gone unpunished; the ALJ had rested on no more than an "implicit assumption" that there must have been more than four speeders among the thousands of derailments. Id. at 2512. The Commission noted that the evidence showed derailments could occur at either high or low speeds and thus concluded that the company's claim that it punished all speeders who caused significant damage was "facially consistent" with the facts. Id. at 2517.

As for the vagueness of a slow order, the Commission noted that the supervisors testified that they thought the employees should have known that such orders implied a ten-miles-per-hour maximum. Id. at 2515. And even conceding the ALJ's conclusion that speedometer tapes and damage estimates were only "roughly accurate" ways to estimate a train's speed, id., the Commission thought this proved only that the company's disciplinary policy on speeding was arguably unfair, not that the company failed to apply the policy consistently. The Commission explained at some length that an ALJ's job was not to enforce good business practices but only to determine whether discipline accorded with a company's normal policy. Id. at 2516-17.

The Secretary petitioned this court for review of the Commission's order under 30 U.S.C. § 816(b) (Supp. V 1981).

II

We hold that the Commission exceeded its statutory authority in reversing the ALJ's findings of fact. To clarify our reasoning, we first establish a proposition that Phelps Dodge denies but that the Commission itself appears to recognize: that the Commission is statutorily bound to uphold an ALJ's factual determinations that are supported by substantial evidence. We then explain why the ALJ's findings in this case were more than amply supported by the record considered as a whole and consequently why the Commission's disagreement with those findings exceeded its statutory standard of review.

A. Commission's Power to Review Factual Findings by an ALJ

Section 113(d) of the Mine Act provides that a person adversely affected by an ALJ's decision may file a petition for discretionary review by the Commission only upon one or more of five specified grounds:

(I) A finding or conclusion of material fact is not supported by substantial evidence.

(II) A necessary legal conclusion is erroneous.

(III) The decision is contrary to law or to the duly promulgated rules or decisions of the Commission.

(IV) A substantial question of law, policy or discretion is involved.

(V) A prejudicial error o[f] procedure was committed.

30 U.S.C. § 823(d)(2)(A)(ii) (Supp. V 1981). The statute then directs that, "[i]f granted, review shall be limited to the questions raised by the petition." Id. § 823(d)(2)(A)(iii). Thus, the only "question" relating to the factual findings of an ALJ that the Commission can consider is whether those findings are supported by substantial evidence.

It is true that the Administrative Procedure Act (APA) empowers agencies to make independent factual findings on the basis of a record developed by an ALJ. See 5 U.S.C. § 557(b) (1976). The Mine Act explicitly recognizes, however, that the "review authority of the Commission" is dictated by section 823(d)(2) of title 30 and not by the contrary section of the APA. 30 U.S.C. § 823(d)(2) (Supp. 1981). The legislative history of the Mine Act suggests the rationale for this scheme. Congress thought it "imperative that the Commission strenuously avoid unnecessary delay in acting upon cases." S.REP. No. 181, 95th Cong., 1st Sess. 48 (1977); see also, e.g., 123 CONG.REC. 4389 (1977), U.S. Code Cong. Admin.News 1977, pp. 3401, 3447 (statement of Sen. Williams) (bill "vastly streamlines nearly every aspect of the enforcement and administration of the mine safety and health program"); id. at 20,048 (statement of Sen. Heinz) (bill "designed to streamline every aspect of the enforcement mechanism"). Limiting the Commission's ability to reopen factual disputes already reasonably settled by an ALJ was apparently one way Congress chose to streamline the enforcement process.

The only wrinkle in the legislative history is that the bill as originally introduced was even more explicit on this point than was the version reported out of committee. The original version declared in what eventually became 30 U.S.C. § 823(d)(2)(C): "The administrative law judge's findings and conclusions of fact, as distinguished from policy determinations, shall not be set aside by the Commission unless such findings or conclusions of fact are unsupported by substantial evidence of record." S. 717, 95th Cong., 1st Sess. § 114(d)(2)(C) (1977), reprinted in SUB-COMM. ON LABOR OF SENATE COMM. ON HUMAN RESOURCES, 95TH CONG., 2D SESS., LEGISLATIVE HISTORY OF THE FEDERAL MINE SAFETY AND HEALTH ACT OF 1977, at 170 (1978). The committee deleted the sentence before it reported the bill to the Senate floor, see id. at 570, though we find no explanation for the deletion in the committee hearings or the accompanying Senate report.

It is apparent, however, that the committee deleted the sentence simply because it repeated what was already stated clearly enough in the remainder of the statute. We have already showed that the final version quite explicitly limits the grounds on which the Commission may grant review. Had the committee intended to alter the Commission's review powers with respect to factual determinations, it would surely have noticed the need to amend these explicit limitations as well. Moreover, the conference committee later changed the placement of the declaration that the Mine Act's provision for Commission review supersedes that portion of the APA giving an agency the right to make independent factual findings. Compare 123 CONG.REC. 20,058 (1977) (S. 717, § 114(d)(2), as passed by Senate) and id. at 23,191 (House insisting on H. 4287, with no comparable provisions) with S.REP. No. 461, 95th Cong., 1st Sess. 28 (1977) (conference version of S. 717, § 113(d)(2)). That the conference committee even briefly focused its attention on the Mine Act's deviation from the APA shows that Congress undoubtedly understood the scheme the Mine Act expressly created.

B. Support for the ALJ's Factual Findings

The Commission appeared to recognize that it was bound to affirm an ALJ's factual determinations supported by substantial evidence: it concluded its opinion with the assertion that "substantial evidence does not support the judge's rejection of Phelps Dodge's defense." Commission decision, 3 F.M.S.H.R.C. at 2517. While voicing an understanding of the statutory standard, however, the Commission did no more than substitute a competing view of the facts for the view the ALJ reasonably reached. There can be little doubt that the ALJ's findings were amply supported by substantial evidence, and even if the Commission's own view found support in the record as well, it was bound to uphold the ALJ's determinations.

The central factual disagreement between the Commission and the ALJ was whether Phelps Dodge would have disciplined Chacon in the same way had he derailed his trains under similar circumstances but not previously fought for safety reforms. All agreed that discipline for infractions like Chacon's was rare, while derailments were quite common. In dispute was whether such discipline was rare because (1) operators only rarely derailed and caused damage when their speedometers indicated excessive speed (with excessive speed defined as greater than ten miles per hour when a general slow order was in effect), or (2) Phelps Dodge did not strictly enforce its speeding rules. The Commission adopted the first view, though it had no affirmative evidence that violations like Chacon's were rare. It fastened instead on the Secretary's lack of direct evidence that speeding was common, criticizing the "implicit assumption" that more than a few of several thousand derailments would cause damage and be attributable to speeding.

We note to begin with that even if the Commission had full authority to find facts on its own, its conclusion is arguably erroneous. The Commission was purporting to require the employer to bear the burden of persuasion on this defense. See id. at 2514 (quoting Pasula, 2 F.M.S.H.R.C. at 2799-800). Although the Secretary may have failed to point to previous violators who went unpunished, Phelps Dodge offered little evidence of its own to show that it consistently enforced its speeding policy. Indeed, the Commission itself concluded that it did "not find this [frequency-of-]derailment evidence particularly persuasive for either party's case." Id. at 2513. Why it then concluded that Phelps Dodge carried a burden of persuasion on its defense is thus not entirely clear.

In any event, considered on the basis of the record as a whole, the ALJ's finding that Phelps Dodge would not have punished Chacon in the same way had he not pressed for safety reforms is supported by substantial evidence. The ALJ had ample reason to infer that the infrequency of discipline for speeding stemmed from a relaxed disciplinary policy rather than an absence of violations like Chacon's. First, there was reason to think that the disciplinary actions the company had previously taken against speeders had been for violations more extreme and more reliably proven than Chacon's. Information about the speeds required to merit one of the rare disciplinary actions was spotty, but the one instance of discipline before Chacon's derailments about which such information was available involved a twenty-miles-per-hour infraction. Moreover, there was evidence that the speedometers were more reliable at speeds above fifteen miles per hour than below.

Second, statements that supervisors made after the derailments occurred support the inference that the company did not usually punish employees in Chacon's position. Lester D. Olson, the mine superintendent who both the Commission and the ALJ agreed had expressed hostility to Chacon's safety activities, said in the company's defense at the grievance proceedings that there had been "[l]ots of derailments" and that the company "had to start somewhere." ALJ decision, 2 F.M.S.H.R.C. at 1279. Obviously, the characterization of Chacon's punishment as a way to begin cracking down on derailments suggests that operators in Chacon's position were not previously subjected to such punishment. Moreover, after Chacon returned from his suspension, the assistant shift foreman on duty during the February 12 derailment told Chacon that, were it left to him, he would not have suspended Chacon and that the discipline had come from the front office. Id. This statement could reasonably support an inference that the foreman, who was presumably familiar with the company's past history of discipline, thought the suspension was unusually harsh.

Finally, there is the significance of the evidence showing the unreliability of the speedometers and the vagueness of a slow order. The Commission seems to have concluded that this evidence could show no more than that the company's policy of discipline for speeders was a poor one, not that the company in fact failed to enforce it. This conclusion, however, is flawed in at least two respects. In the first place, even if this evidence suggested no more than that the speeding enforcement policy would have been unreasonable if strictly adhered to, it is not clear to us why this was not probative on the question of whether Phelps Dodge in fact strictly enforced such a policy. One may generally presume that a company will not enforce a disciplinary policy unreasonable on its face. Because the evidence of the policy's unfairness was plain, it was not erroneous for the ALJ to have found somewhat implausible the assertion that Phelps Dodge applied the policy in all cases. Although this implausibility would do little to cast doubt on positive evidence that the company consistently adhered to the policy, it was probative here, where the evidence showed that Phelps Dodge only rarely punished operators for speeding.

The plainest evidence of the purported policy's unreasonableness was that showing the unreliability of the speedometers. The ALJ could have construed the testimony of one worker to mean that an operator would quite regularly have a speedometer reading of fifteen miles per hour when he was in fact travelling no more than five miles per hour, and vice versa. See J.A. 55. The ALJ could rightly have thought such pronounced unreliability made doubtful the company's assertion that it relied on the tapes from those speedometers as a matter of routine. Only slightly less suspect was the claim that the company regularly enforced an implied ten-miles-per-hour speed limit on slow-order days, even though employees were not aware of any such implicit maximum. To punish employees for violating norms not clearly communicated to them in advance is unfair enough to raise doubt that the company in fact regularly did so. Moreover, the ALJ could have thought it unlikely that a company that does not care enough about a purported rule to make it well known to employees would regularly discipline employees who violate it.

The second flaw in the Commission's reasoning is that the unreliability of the speed tapes and the vagueness of the slow order do not cast doubt on the plausibility of the company's defense simply by showing the unreasonableness of the purported disciplinary policy. In addition, both facts directly undercut the company's claim that only a handful of employees have found themselves in Chacon's position and consequently its claim that this accounts for the infrequency of discipline for speeding. With such erratic speedometers, quite a few of the thousands of trains that derailed in the few years preceding Chacon's discipline must have had speed tapes indicating fifteen miles per hour or more, even if no train in fact ever travelled that fast. It does not seem unreasonable for the ALJ to have thought that the company would have punished far more employees than it did had it punished every operator of a train that derailed on a slow-order day and that bore a high speedometer reading.

By the same token, given that employees were ignorant of any implicit limit, one would expect that they would quite regularly have strayed unknowingly beyond it. Thus, even if the speedometers were completely accurate, the ALJ could reasonably have thought that the company would have punished many more employees than it did had it consistently disciplined all workers in Chacon's position. Indeed, the very fact that there was disagreement over what a slow order implied suggests that the company did not usually enforce the particular meaning of a slow order it attempted to enforce in Chacon's case. For nothing would more quickly have brought an implicit speed limit to the workers' attention than a consistent effort to enforce it.

Considered as a whole, therefore, the record more than amply supported the ALJ's conclusion that Phelps Dodge singled Chacon out for punishment because he had pressed for safety reforms and that it would not have punished him had he derailed his trains in the same manner before he began his safety campaign. Accordingly, the Commission exceeded its statutory authority when it substituted a competing view of the facts for the ALJ's reasonable factual determinations. We note that apparently nothing in the ALJ's factual findings, and certainly nothing in our view of the substantial evidence supporting those findings, hinges on the fact that Pasula required Phelps Dodge to bear a burden of persuasion on its defense rather than a burden of production. We therefore have no occasion to consider attacks by the Secretary and Phelps Dodge on the allocation of burdens of proof established in Pasula, see Boich v. FMSHRC, 704 F.2d 275 (6th Cir. 1983) (striking Pasula rule that employer bears burden of persuasion on defense, rather than mere burden of production).

III

The petition for review is granted and the case remanded to the Commission for disposition consistent with this opinion.

It is so ordered.