The issue in this case is whether the Port Authority of New York and New Jersey may subsidize the operations of its Port Authority Trans-Hudson ("PATH") Railroad with revenues collected from tolls on its bridges and tunnels. The subsidy is achieved by including PATH in the rate base of the tolls that the Port Authority charges on its bridges and tunnels. The question is then whether the bridge tolls are "just and reasonable," as required by section 135(i) of the Federal-Aid Highway Act of 1987, 33 U.S.C. Sec. 508 (Supp. V 1987).

The Automobile Club of New York and five AAA Clubs of New Jersey (the "Auto Clubs") sued the Port Authority after it raised its bridge and tunnel tolls (along with PATH fares) in 1987. The Auto Clubs claimed that, with PATH in the rate base, the new bridge tolls were not just and reasonable. Judge Milton Pollack found that the Port Authority's bridges, tunnels, bus terminal, bus programs and PATH form an "integrated, interdependent transportation system." Automobile Club v. Port Auth., 706 F. Supp. 264, 280 (S.D.N.Y.1989). Accordingly, he decided, it is proper to include PATH in the rate base, from which it follows that the tolls are just and reasonable. Id. at 276-77. We agree and affirm the decision below.

New York and New Jersey created the Port Authority by interstate compact in 1921, N.Y. Unconsol.Law Secs. 6401 et seq. (McKinney 1979),1 with the consent of Congress, Act of Aug. 23, 1921, ch. 77, 42 Stat. 174. The two states recognized that, with the development of commerce, the territory in and around the port of New York had become "commercially one center or district." N.Y.Unconsol.Law Sec. 6401. In creating the Port Authority, they sought to achieve "better co-ordination of the terminal, transportation and other facilities of commerce in, about and through the port of New York." Id.

The Port Authority owns and operates a transportation network which provides river crossings linking southern New York and northern New Jersey. This network consists of four bridges, the George Washington Bridge (including a bus station), the Bayonne Bridge, the Goethals Bridge, and the Outerbridge Crossing; two tunnels, the Lincoln Tunnel and the Holland Tunnel; the bus terminal at 40th Street and 8th Avenue in Manhattan; the Port Authority Bus Programs; and the PATH railroad system. PATH's lines run from Newark and Hoboken, New Jersey to various points in New York's central business district ("CBD," the area of Manhattan south of 59th Street).

The Port Authority's bridges, tunnels and bus terminal, taken together, operate at a profit. For example, in 1986 (before the toll increases at issue here), they produced $243.7 million in revenues and incurred $213.4 million in expenses (including depreciation), yielding $30.3 million in net income. In the same year, PATH lost $92.2 million ($49.0 million in revenues against $141.2 million in total expenses). Thus, the inclusion of PATH (along with $28.8 million in expenses attributable to the depreciation and amortization of the Port Authority's Bus Programs) resulted in a net operating loss that year of $90.7 million for the Port Authority's Interstate Transportation Network. Automobile Club, 706 F. Supp. at 283; Court Exhibit I, Exhibit F.

Tolls on the Port Authority's bridges must be "just and reasonable." This requirement originated in section 4 of the Bridge Act of 1906, 33 U.S.C. Sec. 494 (1982). In 1987, Congress repealed the statutory provision that gave the Secretary of Transportation the power to prescribe "just and reasonable" tolls. Federal-Aid Highway Act of 1987, Pub.L. No. 100-17, Sec. 135(a), 101 Stat. 134, 173 (the "Highway Act"). At the same time, Congress retained the requirement that tolls be "just and reasonable." Id. Sec. 135(i), 101 Stat. at 174 (codified at 33 U.S.C. Sec. 508 (Supp. V 1987)). The House Conference Report accompanying the Highway Act explained that the "just and reasonable" requirement

is not intended to interfere with or in any way restrict existing authority granted to multi-modal transportation agencies, such as the Port Authority of New York and New Jersey, that operate bridges along with other facilities to use bridge toll revenues for nonbridge purposes or the pooling or combination of the revenues from all of their facilities.

H.R.Conf.Rep. No. 27, 100th Cong., 1st Sess. 175, reprinted in 1987 U.S.Code Cong. & Admin.News 66, 159.

In 1987, the Port Authority decided to engage in a $1.5 billion capital improvement program. Capital spending was necessary in order to maintain and modernize its interstate transportation network as use of its facilities increased rapidly. The Port Authority planned to spend $830 million to improve the PATH system, $495 million on Port Authority bridges, tunnels and bus terminals, and $150 million to establish a ferry service between Hoboken and lower Manhattan. The existing level of tolls and fares would be inadequate to operate the transportation network and to finance this capital program. Automobile Club, 706 F. Supp. at 281. The Port Authority therefore increased the tolls on its bridges and tunnels, as well as the fares for travel on PATH. For example, non-discounted, eastbound--no toll is charged for vehicles traveling west--passenger car tolls on bridges and tunnels rose from $2.00 to $3.00, and the price of twenty commuter discount tickets to pay such tolls rose from $20.00 to $40.00. PATH fares, which are charged for travel in each direction, were increased from $.75 to $1.00. Id. at 267-68.

The Auto Clubs are non-profit organizations which promote the interests and welfare of motor vehicle users. They sued for a declaratory judgment that the 1987 increase in the Port Authority's bridge tolls violated the Highway Act's requirement that tolls be just and reasonable. See 33 U.S.C. Sec. 508 (Supp. V. 1987). They also alleged that the Port Authority's action violated 42 U.S.C. Sec. 1983 (1982) and the Commerce Clause of the Constitution. The Auto Clubs claim that PATH should not be included in the rate base of the Port Authority's bridges and tunnels because PATH is unrelated to the other river crossings and because its fares remain too low, even after the 1987 increase. They then assert that, if PATH is excluded from the rate base, the Port Authority's rate of return on its unamortized investment in use is excessive. The Auto Clubs do concede, however, that the rate of return is reasonable if the replacement cost method of accounting is used. In addition, they accept that the rate of return is reasonable, regardless of the method of accounting used, if PATH is properly included in the rate base. Automobile Club, 706 F. Supp. at 268-69 & n. 3. Thus, the principal issue in their lawsuit was the propriety of including PATH in the rate base.

Most of the relevant facts were stipulated by the parties; Judge Pollack received the evidence and heard testimony from witnesses during a two-day bench trial. In his decision, he observed that the legislative histories of the Bridge Act and the Highway Act shed little light on the "just and reasonable" requirement. The 1987 Conference Report did, however, indicate sufficient "flexibility" to permit the Port Authority to include facilities other than bridges, tunnels and terminals in its rate base. Id. at 273.

Judge Pollack also canvassed the case law. He found that City of Burlington v. Turner, 336 F. Supp. 594 (S.D. Iowa 1972), modified and aff'd, 471 F.2d 120 (8th Cir.1973), simply established that the "just and reasonable" rule must include a reasonable return on the investment of capital. Automobile Club, 706 F. Supp. at 273 n. 9. The case of Clarksburg-Columbus Short Route Bridge Co. v. Woodring, 89 F.2d 788 (D.C.Cir.), remanded for dismissal as moot, 302 U.S. 658, 58 S. Ct. 365, 82 L. Ed. 509 (1937), had greater relevance: it involved two competing bridges. The Secretary of War (exercising the power later entrusted to the Secretary of Transportation) ordered the tolls on one bridge lowered. The court held that the Secretary erred in considering only the bridge owner's return on his investment. Competitive factors should also be considered, as in I.C.C. ratemaking. The other bridge might be driven out of business by the reduced tolls; therefore, the owner of that bridge had a right to be heard before the tolls were lowered. 89 F.2d at 793-94. Clarksburg-Columbus thus "established the principle that a toll producing a seemingly unreasonably high rate of return may in fact be reasonable if its rate base does not include other facilities which are related to the facility in question." Automobile Club, 706 F. Supp. at 273. In this case, we are concerned with different facilities that are owned and operated by one entity, the Port Authority. The principle in Clarksburg-Columbus would, mutatis mutandis, support the inclusion of those facilities in a single rate base if they are related to each other.

Finally, there was Automobile Club v. Cox, 592 F.2d 658 (2d Cir.1979), a prior incarnation of the present case. In Cox, the Auto Clubs contested the decision of the Federal Highway Administrator to approve increases in the Port Authority's bridge tolls that went into effect in 1975. The Auto Clubs adopted an "all or nothing" strategy in the lawsuit: They resisted the inclusion of any other river crossing facility--i.e., not just PATH but the tunnels and the bus terminal, as well--in the rate base of the four bridges. The underlying factual issue thus was whether PATH was sufficiently related to the Port Authority's bridges to justify subsidizing PATH with revenues from bridge tolls. The court reviewed that action under the "arbitrary [and] capricious" standard of 5 U.S.C. Sec. 706(2)(A), since the administrative approval of toll rates for the future constitutes rulemaking. Id. at 664. Writing for a unanimous panel of this court, Judge Henry J. Friendly sustained the action of the Administrator because it was not arbitrary or capricious. Id. at 673.

Judge Pollack looked beyond this "sparse" case law construing the Highway Act's "just and reasonable" requirement, Automobile Club, 706 F. Supp. at 273, to consider public utility ratemaking cases where the same requirement is applied to gas and electricity rates, id. at 274. In those cases, courts apply the "used and useful" principle to decide whether different facilities may be included in one rate base. Id. at 274-75 (quoting Tennessee Gas Pipeline Co. v. FERC, 606 F.2d 1094, 1109 (D.C.Cir.1979), cert. denied, 445 U.S. 920, 100 S. Ct. 1284, 63 L. Ed. 2d 605, 447 U.S. 922, 100 S. Ct. 3012, 65 L. Ed. 2d 1113 (1980)); see also Smyth v. Ames, 169 U.S. 466, 546, 18 S. Ct. 418, 433, 42 L. Ed. 819 (1898). In reviewing utility rates, courts take into consideration the cost of efforts to improve the quality and quantity of service "systemwide." E.g., Natural Gas Pipeline Co. v. FERC, 765 F.2d 1155, 1164 (D.C.Cir.1985), cert. denied, 474 U.S. 1056, 106 S. Ct. 794, 88 L. Ed. 2d 771 (1986).

The utility ratemaking cases "suggest by analogy" that "Port Authority activities which represent efforts to improve the quality and quantity of its services systemwide ... may be included in the Port Authority's rate base." Automobile Club, 706 F. Supp. at 276. Judge Pollack also concluded that a "systemwide" approach to this case precludes the use of a narrow "origin and destination" analysis to decide whether all of the Port Authority facilities at issue are sufficiently related to each other to justify including them in the same rate base. Id.

Applying these principles, Judge Pollack held that it was proper to include PATH in the same rate base with the Port Authority's bridges and tunnels: There is a "spillover" effect by which an interruption of service on PATH would disrupt the flow of traffic on the bridges and tunnels. The existence and smooth functioning of PATH therefore benefits those other facilities, and they would suffer serious congestion if PATH ceased to exist. The "spillover" effect means that PATH is "related" to the bridges and tunnels, notwithstanding the absence of a substantial "origin and destination" relationship between PATH and some of the bridges. See id. at 269-70, 276. Judge Pollack also found that users of PATH pay their fair share of the expenses of crossing the Hudson River. Id. at 277. Thus, with PATH properly included in the rate base, the Port Authority's bridge and tunnel tolls were held to be just and reasonable. Id.2

The standard of our review is a threshold issue. Appellants of course urge plenary review of the decision below; appellee points out that the district court's legal conclusions rest upon factual findings that must be accepted unless they are clearly erroneous.

Judge Pollack based his decision upon several legal conclusions. For example, he decided that it was not necessary to use a narrow "origin and destination" analysis to show that PATH is related to the Port Authority's bridges and tunnels; the "spillover" effect could also show the requisite relationship. He then applied that rule to the facts of this case and concluded that PATH is properly included in the rate base with the bridges and tunnels. It followed from that conclusion that the tolls are "just and reasonable," as required by the Highway Act.3 These legal conclusions are subject to plenary review in this court. See Karavos Compania Naviera S.A. v. Atlantica Export Corp., 588 F.2d 1, 8 (2d Cir.1978) (application of a legal standard is not a question of fact within Fed.R.Civ.P. 52(a)).

Judge Pollack's legal conclusions rested upon numerous findings of fact which, under Fed.R.Civ.P. 52(a), we must accept unless they are clearly erroneous. Among the most important of those findings was the fact that closing one of the Port Authority's facilities, for example PATH, would increase congestion on the others--the "spillover" effect. Automobile Club, 706 F. Supp. at 279-80. This contributed to his factual finding that the Port Authority's bridges, tunnels, bus terminal, bus programs and PATH form "an integrated interdependent transportation system." Id. at 280.

Thus, our standard of review depends upon what we are reviewing. It is a factual finding, subject to Rule 52(a)'s "clearly erroneous" requirement, to decide whether a "spillover" effect exists. It is then a legal conclusion, subject to plenary review, to decide what relevance any "spillover" effect might have--i.e., whether it shows that PATH is related to the bridges and tunnels. Cf. Patel v. Wargo, 803 F.2d 632, 634 (11th Cir.1986) (whether company was an employer under Fair Labor Standards Act was a legal determination, but "individual findings of fact which led to that legal determination must be examined under the clearly erroneous standard").

The dispositive issue in this case is whether there is a close relationship between PATH and the Port Authority's bridges and tunnels. Both sides agree that PATH must be "related" to the bridges and tunnels if it is to be included in their rate base. They disagree over how to define what they refer to as a "functional relationship" between river crossings, how to identify a functional relationship and how to measure it.

The Auto Clubs contend that the relationship must be "direct and significant" and that this can only be shown by "origin and destination" analysis. A functional relationship based upon "origin and destination" analysis means that one river crossing substitutes relatively easily for another. If a large percentage of the people traveling between A in New Jersey and B in New York freely choose to use either river crossing X or river crossing Y, then X and Y have a functional relationship that the Auto Clubs consider direct and significant. Using this "origin and destination" analysis, the Auto Clubs concede that there is a relationship between PATH and the Holland and Lincoln tunnels, but argue that PATH has only "some small" functional relationship with the George Washington Bridge and practically no functional relationship with the three Staten Island bridges.4

The Port Authority acknowledges the importance of "origin and destination" analysis. It carries out "origin and destination" studies of its own and relies upon them for its planning. The Port Authority does not claim that "origin and destination" analysis would show that PATH is related to all of the other river crossings. However, it insists that "origin and destination" analysis alone is too narrow for the complete description or measurement of a "functional relationship." Another component of a functional relationship, the Port Authority argues, is the "spillover" effect.

According to the Port Authority, each of its river crossing facilities will suffer if any one of them is eliminated; this results from the "spillover" (or "domino") effect. The reason for this is that use of the river crossings is now close to capacity during peak morning hours. The Port Authority does not claim that, without PATH, significant numbers of PATH riders would travel as far afield as the Outerbridge Crossing. Clearly, the Holland and Lincoln tunnels are the most direct alternatives to PATH. However, nearly 70,000 passengers use PATH during the peak morning hours. If they were to use the tunnels instead, the tunnels would become overcrowded. Congestion in the tunnels would lead many travelers (including some who normally use PATH) to use the George Washington Bridge. Others (for example, those whose destination is Long Island, rather than the CBD) could shift to the Staten Island bridges. Indeed, there is even a significant "origin and destination" relationship between the George Washington Bridge and the Outerbridge Crossing: The former serves as a northern bypass of the CBD, while the latter is a southern bypass. If, then, the George Washington Bridge is congested, some people traveling between New Jersey and Long Island will choose instead to use the Outerbridge Crossing. Congestion on one Staten Island bridge will divert traffic to the other Staten Island bridges. Thus, PATH benefits the entire interstate transportation network because the effects of eliminating PATH would extend all the way north to the George Washington Bridge and south to each of the Staten Island bridges. In short, concludes the Port Authority, the "spillover" effect demonstrates that there is a functional relationship between PATH and the other river crossings.

We agree with the position taken by the Port Authority and accepted by the district court. Measuring the relationship between river crossings by means of "origin and destination" data is useful but not sufficient. This approach is fine when all lanes are open and traffic flows smoothly, but it does not adequately describe what happens when traffic is disrupted. It fails to take into account the full extent of the Port Authority's responsibilities. As anyone who has ever sat in a line of traffic or a railroad station for a length of time is aware, the Port Authority's employees cannot think only in terms of traffic flowing smoothly and trains running on time; they must also anticipate and respond to disruptions and delays. The Port Authority is entitled to prove that PATH is related to the bridges and tunnels with evidence that a "spillover" effect links even those river crossings that travelers do not normally use interchangeably. A functional relationship exists, then, among those facilities that contribute to the Port Authority's performance of its duty efficiently to provide transportation to travelers over and under the waterways between southern New York and northern New Jersey. See Automobile Club, 706 F. Supp. at 276.5 Thus, Judge Pollack was correct in following the analogy of the utility ratemaking cases and in deciding not to measure PATH's relationship to the other river crossings narrowly in terms of "origin and destination" data alone. We agree that, as a matter of law, the "spillover" effect should also be considered.

Judge Pollack found that direct relationships exist between different river crossing facilities--for example, between PATH and the two tunnels, and between the Holland Tunnel and the Staten Island Bridges. He found that closing one or more of those facilities increases congestion in the others--the "spillover" effect. Evidence also established that it was most efficient and cost-effective to operate the Port Authority's interstate transportation facilities as an integrated system. These and other factual findings formed the basis for Judge Pollack's finding that the Port Authority's bridges, tunnels, bus terminal, bus programs and PATH constitute an "integrated, interdependent transportation system." See Automobile Club, 706 F. Supp. at 279-80.

Our review of the record convinces us that these factual findings are not clearly erroneous; indeed, it requires us to agree with Judge Pollack's findings. The evidence of increased use of the Port Authority's facilities is particularly significant. The number of passengers traveling east via those facilities by auto, bus and rail during the morning peak period (7:00-10:00 a.m.) has increased since 1967 as follows:

             1967    1977    1980    1987
Automobile  55,200  66,900  69,000  78,500
Bus         75,100  63,400  69,500  81,500
PATH        43,700  48,200  50,800  69,400

Id. at 267; Defendants' Exhibit F. The evidence in this case indicates that the Port Authority's facilities now operate at the limit of their capacity during peak periods. If PATH did not exist, the influx of nearly 70,000 passengers during each morning peak period would severely disrupt traffic on the bridges and tunnels, even with the additional carpooling that would inevitably occur. If PATH had to be replaced, it would annually cost more to do so than the current subsidy. Id. at 280. PATH thus benefits those who normally use the Port Authority's bridges and tunnels. These facts lead us to agree with the legal conclusion that there is a functional relationship between PATH and those bridges and tunnels; this justifies including PATH in the rate base.

This conclusion is consistent with the result in Cox. However, one need only read Judge Friendly's opinion to recognize that we must, to a certain extent, distinguish Cox in order to justify today's decision in favor of the Port Authority.

Cox offered hope to both sides in this case. It did, of course, sustain a fare and toll structure that subsidized PATH. On the other hand, Judge Friendly wrote with palpable reluctance. The inclusion of PATH in the rate base was a "giant leap" from anything previously acceptable under the "just and reasonable" requirement. Cox, 592 F.2d at 672. He suggested that the deferential "arbitrary or capricious" standard of review required the result. He noted that the Auto Clubs' "all or nothing" strategy of seeking to limit the rate base to the four bridges was flawed, and he invited them to submit a "new protest." Id. at 673. Consequently, the Port Authority handles Cox gingerly; it strives to distinguish the case that it won a decade ago. By the same token, the Auto Clubs embrace Cox, confident that it now dictates a contrary result, especially since they have abandoned their previous "all or nothing" strategy.

The Auto Clubs make much of Judge Friendly's concern about taking the "giant leap" of including PATH in the rate base. However, as Judge Pollack pointed out, the Cox court did, in fact, take that leap; and the real issue is whether the leap is proper, not how long it is.

We find it easier to sustain the Port Authority's position today than the panel did ten years ago. What was a difficult leap for one of the giants of our jurisprudence has become a comfortable step for us--not because we presume to match Judge Friendly's stride, but because the situation in 1989 differs from that of the mid-1970's. There is now considerably more evidence in the record to show that PATH is related to the other river crossing facilities. See generally Automobile Club, 706 F. Supp. at 266-71, 278-83. The use of the Port Authority's facilities has greatly increased since the record in Cox was established, in part because of New York's and New Jersey's economic growth in the early 1980s. Evidence before the Cox court showed that 174,800 people used the Port Authority's facilities to travel east across the Hudson during the morning peak period. (An additional 13,500 used Conrail to cross the Hudson.) Cox, 592 F.2d at 671. In 1987, 229,400 passengers used Port Authority facilities to travel east each morning peak period, with an additional 24,000 using other rail services into Penn Station. Automobile Club, 706 F. Supp. at 267. The Auto Clubs have not refuted the Port Authority's evidence that its river crossing facilities are now strained to their limits during those peak periods. (We suspect that Auto Club members who commute between New Jersey and New York are ruefully aware of this situation.) The lack of spare capacity during peak periods means that the "spillover" effect must be considered when assessing relationships between the Port Authority's river crossings.

The Auto Clubs point out that the interstate transportation network did not break down completely when service on PATH was interrupted by lengthy strikes in 1973 and 1980. This is evidence, they say, that refutes the Port Authority's claims about the "spillover" effect. We disagree. What happened in 1980 during a temporary interruption of service tells us little about what would happen now if PATH were to disappear permanently. The number of people using the Port Authority's river crossings during morning peak hours increased by 22% between 1980 and 1987. See supra at 423. The Auto Clubs' argument does not take into consideration the current absence of spare capacity on the bridges and tunnels to absorb displaced PATH riders. In addition, testimony at trial showed that the Port Authority made special arrangements to alleviate the congestion caused by the 1980 strike, and commuters adjusted their schedules to avoid traffic jams on the bridges and tunnels. Some of these stopgap measures would be more difficult to take now, and few of them would be feasible over the long run if PATH ceased to exist.

In Cox, Judge Friendly looked askance at PATH's fares, which had not been increased in 1975 when the bridge and tunnel tolls were raised. See Cox, 592 F.2d at 672 (30 cent fare is "prima facie unreasonably low"). The Auto Clubs pursue that point, decrying the Port Authority's failure to conduct a "formal study" before raising PATH's fares in 1987 and arguing that PATH users do not pay their fair share of the cost of traveling across the Hudson.

PATH fares have been increased three times since this court's decision in Cox. See Automobile Club, 706 F. Supp. at 281. The most recent increase--from $.75 to $1.00--coincided with the toll increases at issue in this case. PATH fares are now quite comparable to bridge and tunnels tolls. Indeed, it could be argued that it is now the motorists who get a better deal. A round-trip on PATH costs $2.00, exactly the cost of a round-trip on the bridges and tunnels using a commuter discount ticket. And, of course, the cost of a bridge or tunnel toll can be shared among all of the riders in a car, whereas each PATH rider must pay his or her full fare.

The Auto Clubs nevertheless suggest that the Port Authority could and should have pushed PATH fares even higher. Again, we disagree. The Port Authority considered the appropriate factors when it set the new fares in 1987--for example, the amount of money that the increases would bring in, comparable fares on New York's subways, the convenience of collecting the fares, and acceptance of the new fares among passengers. See Automobile Club, 706 F. Supp. at 268, 276-77. It was not necessary for the Port Authority to conduct a more formal study or to test the market's elasticity by pushing PATH fares further upward. In short, PATH's fares are no longer "prima facie unreasonably low"; PATH's riders are paying their fair share of the cost of their transportation across the Hudson.

Judge Friendly contemplated the possibility that Congress might "remove or alter the points in controversy." Cox, 592 F.2d at 673. This has not occurred. However, the House Conference Report accompanying the 1987 Highway Act recognized that the Port Authority has authority "to use bridge toll revenues for non-bridge purposes." H.R.Conf.Rep. No. 27, 100th Cong., 1st Sess. 175, reprinted in 1987 Code Cong. & Admin.News 159. We agree with Judge Pollack that this gives the Port Authority some flexibility which could extend to including PATH in the rate base with the bridges and tunnels. See Automobile Club, 706 F. Supp. at 273.

To summarize on this point, we recognize that Cox alone does not provide strong support for holding that the Port Authority's 1987 tolls are just and reasonable. However, the Auto Clubs are mistaken in arguing that there is some rule buried in Cox that requires a different result this time. This case has indeed changed during the past ten years, but in the opposite direction; the facts in the record now make it easier for us to conclude that PATH should be included in the rate base.

PATH, then, is related to the Port Authority's bridges and tunnels. It is proper to include PATH in their rate base, and the tolls are just and reasonable. Judgment affirmed. Judge Van Graafeiland dissenting in a separate opinion.

Because cases decided by this Court often return to us like the proverbial bad penny, we have established the practice, whenever feasible, of assigning the returned case to a panel which includes at least one member of the original panel. This sensible practice insures that at least one member of the second panel will be personally familiar with the issues and arguments presented to the first and with the reasoning that led to the original holding. Eleven years ago, I sat on a panel with two of the strongest members of this Court, Judges Friendly and Timbers. We listened to the identical arguments that the Port Authority made to the present panel, and we rejected them.

Principal among those arguments was PATH's alleged role in relieving congestion on all the bridges, including the Staten Island and George Washington Bridges. The Federal Highway Administrator, whose decision was being appealed, held that PATH was "used and useful" for the convenience of both passenger vehicles and trucks using the Staten Island and George Washington Bridges. Our response to these statements was as follows:

The question was not whether PATH requires a subsidization but whether, in the absence of Congressional action such as the 1952 legislation concerning DRPA, the huge deficits of PATH can properly be loaded onto users of the PA bridges--three of which have almost no functional relationship and the fourth, the George Washington Bridge, only a small one. There was no showing that it was the level of the previous tolls rather than the convenience of the bridges and tunnels that was drawing traffic away from PATH or that the higher ones would increase PATH ridership. The higher bridge tolls are paid not simply by commuters who opt to use the George Washington Bridge when they could as readily have used PATH but by commuters for whom PATH was not a feasible alternative, by non-commuter passengers including many who were not destined to or did not originate in Manhattan or nearby New Jersey, and even by trucks.

592 F.2d at 672.

The Federal Highway Administrator had no problem thereafter in interpreting correctly our 1979 decision:

The Circuit Court's criticism of the Administrator's prior decision in 1976 including PATH in the rate base is well taken.... If the Port Authority were allowed a revenue requirement based on its total activities or on its total surface transportation facilities, it would be free to invest in deficit operations secure in the knowledge that they could charge the bridge users for support of such activities. A finding along those lines would be contrary to the statutory duty to protect the bridge users' interests and would be an abdication of statutory responsibilities.

* * *As a matter of law, the rate base for the bridge tolls is limited to the four bridges, two tunnels, and the bus terminal.

Investigation Report and Recommendations, F.H.W.A. Dkt. No. 841-T, 15-16, 18 (June 14, 1985), adopted and incorporated by reference, Order of the Federal Highway Administrator, F.H.W.A. Dkt. No. 841-T, 2 (June 19, 1985).

I respectfully disagree with my present colleagues' attempts to distinguish our 1979 decision. Although I do not profess to have greater legal acumen than they, I have something that they lack--firsthand knowledge of what transpired on the prior appeal. With the benefit of this background, I am convinced that there is no material difference between the 1979 case and the instant case. In short, the instant appeal is a classic example of deja vu. Accordingly, because I am unable to reconcile my colleagues' present holding with our 1979 decision, I dissent. However, I want to make it clear that, even if I were approaching this matter de novo, I would conclude, as I did before, that it is manifestly unfair and improper to saddle captive bridge users with the enormous deficits of PATH.

One need glance only briefly at a map of the United States to see that New York City is the gateway to all of Long Island and southern New England for interstate vehicular traffic from the west and south. The Port Authority, a statutory monopoly, N.Y.Unconsol.Laws Sec. 6513 (McKinney 1979), holds the key to the gate. This means that a motorist taking the most direct route from a city such as Pittsburgh, Pennsylvania to New Haven, Connecticut must pay the Port Authority a fee for the use of that key; so also must a motorist going from a city such as Washington, D.C. to Riverhead, Long Island. The most convenient passageways through the gate for such motorists would be the George Washington Bridge and the Staten Island bridges.

The Port Authority itself describes the George Washington Bridge as a "key link in the northern metropolitan bypass system" which provides connections between Route 80, the main interstate highway across New Jersey, and the New York Thruway, "as well as other regional highway systems in each state." Upon closer observation, the aforesaid map would show also that Yonkers, White Plains, Mount Vernon, New Rochelle, and numerous other population enclaves all are located immediately north of the George Washington Bridge, which serves as their interstate gateway to and from the south and west. While approximately fifty million vehicles cross the George Washington Bridge in an easterly direction each year, it is undisputed that the bulk of this traffic is not headed for the central business district of Manhattan, which is the district served by PATH.

My colleagues belittle this undisputed "origin and destination" evidence concerning the bridge traffic and emphasize the so-called "spillover" or "domino" effect that they say would result if PATH were eliminated. Apparently, my colleagues believe that this "spillover" effect was not considered by the prior panel. If this is what they believe, they are in error:

The PA witness stated that, because of its location, the George Washington Bridge traffic is "predominately peripheral" (Ex. PA 7, p. 25); shutting down PATH thus would affect it in the first instance less than the tunnels; it is the overcrowding of the latter that might congest the Bridge.

592 F.2d at 672.

The prior panel was well aware of the off-bridge traffic problems that a motorist would encounter should he or she decide to participate in a "spillover" from the two tunnels to the George Washington Bridge. It is approximately ten, heavily-traveled miles from the west end of the Lincoln Tunnel to the George Washington Bridge if the motorist elects to go by way of the tolled New Jersey Turnpike. It is approximately eight equally-crowded miles if the motorist proceeds along secondary roads. From the Holland Tunnel to the George Washington Bridge, the distance via the Turnpike is approximately thirteen miles, and, via secondary roads, approximately eleven miles. From the east end of the George Washington Bridge, the distance along the most direct city streets to the east end of the Lincoln Tunnel is approximately seven miles; via the FDR Drive, it is approximately ten miles. From the east end of the Bridge to the east end of the Holland Tunnel, a motorist traversing city streets will drive about ten miles before reaching the Holland Tunnel. If the motorist proceeds by way of the FDR Drive, he or she will travel about thirteen miles.

I was prepared in 1979 to take judicial notice of the fact that travel conditions on the FDR Drive during morning and evening rush hours are extremely bad, and I continue to be so prepared. In my opinion, few, if any, intelligent motorists headed for the central business district of Manhattan will drop out of line at either the Holland or Lincoln Tunnels during the morning rush hours, detour some fifteen to twenty miles to cross the George Washington Bridge and join the morning traffic from the Triborough Bridge crawling south on the FDR Drive. Travel along the east bank of the Hudson River would be no better than on the FDR Drive, as witness the recent much-litigated, but unsuccessful, effort to construct the "Westway." See Sierra Club v. United States Army Corps of Engineers, 776 F.2d 383, 386 (2d Cir.1985), cert. denied, 475 U.S. 1084, 106 S. Ct. 1464, 89 L. Ed. 2d 720 (1986).

The relationship between the Staten Island bridges and PATH is equally remote. A survey by the Port Authority of the traffic patterns on these bridges states that they form "the southern route ... around the congested Manhattan Central Business District" and concludes:

[U]nlike the facilities which feed the Manhattan Central Business District, transit plays virtually no role in this market.

Port Authority of New York and New Jersey, Staten Island Bridges Study 4, 9 (1987).

I am unpersuaded by my colleagues' "spillover" argument that travelers whose destination is Long Island rather than the central business district of Manhattan might shift from the tunnels to the Staten Island bridges. This argument assumes that a substantial number of motorists using the two tunnels during rush hours are headed for Long Island rather than the central business district, an assumption without support in the record and equally unsupportable as a practical proposition.

Treatment of the Port Authority's bridges, tunnels, bus terminal, bus programs and PATH as "an integrated interdependent transportation system", takes no cognizance of the fact that there are several other transit systems available for use by New Jersey residents, the New Jersey Transit and Amtrak trains running from Newark to the Pennsylvania Station and the New Jersey Transit trains running from Newark to Hoboken. It also overlooks the fact that there are trans-Hudson ferries, and that, when the district court's opinion was written, two new ferries were scheduled to begin operations in the then near future, and are now operating. The argument that the closure of PATH will force all commuters to transfer to the tunnels and bridges just doesn't wash.

My final observation on the "spillover" effect is that it hypothesizes a crisis that in fact does not exist, a crisis predicated on the erroneous proposition that the tolls paid by captive users of the bridges and tunnels are the only proper source of additional funds for PATH's needs. I suggest that, before taking up the question how much the largely unaffected users of the George Washington Bridge and the Staten Island bridges can be expected to contribute to the operation of PATH, both the Port Authority and the courts should make a thorough and fair inquiry into whether PATH is contributing its fair share of the Port Authority's earnings, see The Minnesota Rate Cases, 230 U.S. 352, 434-35, 33 S. Ct. 729, 754-55, 57 L. Ed. 1511 (1913); Smyth v. Ames, 169 U.S. 466, 541, 18 S. Ct. 418, 432, 42 L. Ed. 819 (1898), and how much the users of PATH should be expected to pay to accomplish this result. See Hughes v. Oklahoma, 441 U.S. 322, 336, 99 S. Ct. 1727, 1736, 60 L. Ed. 2d 250 (1979).

On several occasions, the Federal Highway Administrator, whose decision we reviewed in 1979, expressed a concern that bridge users should not be saddled with exorbitant tolls for the purpose of subsidizing the users of PATH:

As the legislative history of the General Bridge Act makes clear, one guiding purpose of the statute was to prevent the operators of toll bridges from exploiting what was often a monopoly position to exact excessively high tolls from bridge users.

Decision of Federal Highway Administrator, F.H.W.A. Dkt. No. 76-9, 5 (August 9, 1977).

In this regard, it should be repeated that while a fair rate of return is an important element of financial need in a bridge toll proceeding such as this, the burden of maintaining such a rate cannot be allowed to fall exclusively on users of the subject bridges.

Decision of Federal Highway Administrator, F.H.W.A. Dkt. No. 76-9, 12 (November 7, 1977). See also The Minnesota Rate Cases, supra, where the Court said that "the carrier cannot justify unreasonably high rates on domestic business because only in that way is it able to meet losses on its interstate business." 230 U.S. at 435, 33 S.Ct. at 755.

Because any future toll increase must be passed upon by the Administrator, adequate safeguards exist to prevent discrimination against bridge users in favor of mass transportation users.

F.H.W.A. decision of November 7, 1977, supra, at 12.

When Congress transferred the exercise of these safeguards to the district courts, it surely anticipated that the courts would not approve the Authority's claim of reasonableness without receiving and passing upon facts and figures to support it. See Lincoln Gas & Electric Light Co. v. Lincoln, 223 U.S. 349, 361-65, 32 S. Ct. 271, 273-75, 56 L. Ed. 466 (1912); Chicago, Milwaukee & St. Paul Ry. v. Tompkins, 176 U.S. 167, 179-80, 20 S. Ct. 336, 340-41, 44 L. Ed. 417 (1900). I regret to say that I do not believe that has occurred in the instant case. I find nothing in the record that justifies PATH's fare of $1.00, a fare that pays about half the expense of running the railroad and has caused an increase in the PATH deficit from $34 million in 1977 to $87 million in 1987, and to a projected 1991 deficit of $133 million.

The district court correctly found that the Port Authority did not conduct any formal studies on the effect of various PATH fare levels on PATH ridership. 706 F. Supp. at 277. Unlike my colleagues, I do not believe that the district court's finding that the "PATH Board 'knew' that every ten-cent increase in PATH fares produces about $6 million in revenues if ridership remains constant" indicates a thoughtful analysis of possible fare increases. Id. A third-grade student could multiply PATH's sixty million annual passenger figure by ten cents and come up with a figure of $6 million. The same student could determine even more quickly that, if ridership remained constant, an increase of $1.00 in PATH fares would produce $60 million in increased revenues. The decision-making process involved in fare increases does not follow such a simple course.

It is generally accepted that a substantial increase in the cost of public transportation will create what is known as a "resistance factor." See Powell v. Washington Metropolitan Area Transit Comm'n, 485 F.2d 1080, 1083 n. 17 (D.C.Cir.1973). What a reviewing court needs to know, therefore, is what effect the resistance factor would have on any proposed fare increase. We know here that PATH fares have been increased on several occasions and its trains still are running at capacity. No facts, figures, or projections have been introduced into the record to indicate that PATH would not continue to operate at capacity if its fares were increased:

Q. Did you or anyone under your direction prepare any study to determine what the loss in ridership would be under various permutations and combinations of the toll increase, assuming the toll increase was 25 cents to a dollar at that time, assuming it was 50 cents to a $1.25, assuming that it was 75 cents to a $1.50, did anyone prepare any studies as to what the expected ridership would be under the various permutations and combinations?

A. Not to my knowledge.

* * *

Q. Did you make any study as to whether if you had increased the PATH fare to $1.50 you would lose any substantial amount of ridership?

A. No study was made of that.

Testimony of John Haupert, Port Authority Treasurer, Joint Appendix at 265-67.

The reliance placed by the district court and my colleagues upon the "administrative convenience of charging 'round figures' for its tolls and fares" is mystifying to me. This "administrative convenience" argument was explained by Mr. Haupert as follows:

On the PATH, what I was trying to point out to you was that we were going to a four quarter system, which meant that people would have to insert four quarters into the turnstiles. If we went to a five quarter system or a six quarter system it would slow down the amount of people we could get through the turnstiles which would have been an added inconvenience on the passengers. We did reflect that in our thinking as to what would be the appropriate fare.

Joint Appendix at 263.

Undoubtedly, it would take a second or two longer for a passenger to deposit six quarters in a slot than it would take to deposit four. However, to use this as a justification for increasing tolls on the George Washington Bridge some $50 million a year borders on the absurd. Whatever minimal delay might result from the insertion of two more quarters readily could be compensated for by installing an additional turnstile at each station.

If the Port Authority wishes to encourage the use of dollar bills, a fare of $1.50 would require only one bill and two quarters, and a fare of $2.00 would require only one additional bill. Such increases, which easily could be handled without material delay, would help to reduce the $87 million deficit which the Port Authority has placed on the backs of captive bridge motorists and truckers.

I am equally mystified by the district court's statement that PATH's tolls are in line with "other urban subway transit systems", 706 F. Supp. at 277, a finding upon which my colleagues rely in holding that a comparison of fares with the New York City subways is an "appropriate factor". PATH is not an "urban subway transit system". It is an interstate "Railroad", formerly known as the Hudson and Manhattan Railroad Company, and, like other railroads, it operates "Trains". Neither the Port Authority nor the district court has pointed to any similarities between PATH and the New York subway system that justify such a comparison. The Port Authority's proof with regard to the New York subway system can be summed up in one sentence:

Well, we looked at the New York City fare of one dollar and thought that a comparable fare on PATH would be readily acceptable.

Testimony of John Haupert, Joint Appendix at 262.

Admittedly, one of the most frequently used methods of testing the reasonableness of a rate is to compare it with rates on similar traffic in adjacent territories. See Aluminum Co. of America v. I.C.C., 581 F.2d 1004, 1008 (D.C.Cir.1978); Alldredge, Rate-Making for Common Carriers, 94 (1929); 13 C.J.S. Carriers Sec. 297f at 678-80 (1939). However, comparisons of rates have little probative value unless the conditions of operation are essentially the same. Aluminum Co., supra, 581 F.2d at 1008; Chicago & Eastern Illinois Ry. v. Commerce Comm'n, 343 Ill. 117, 125, 175 N.E. 8 (1931); 13 C.J.S., supra, at 679-80; 64 Am.Jur.2d Public Utilities Sec. 213 at 720 (1972). In contrast to the lack of similarity between PATH and the New York City subway system, there are a number of other commuter railroads serving New York City whose services are very similar to those furnished by PATH. I take judicial notice of both their existence and their published fares.

PATH is a commuter train that runs from Newark, a major New Jersey city, to downtown Manhattan. En route it passes through Harrison and Jersey City. It also has a separate New York feeder running from Journal Square in Jersey City to Hoboken, New Jersey and through it to Manhattan. New Jersey Transit and Amtrak also have lines running from Newark to Manhattan. The fare for this service on New Jersey Transit is $2.25, on Amtrak it is $5.75. New Jersey Transit trains also run from Newark to Hoboken. The fare for this service is $1.75, to which must be added the cost of transportation across the Hudson River. The fare by New Jersey Transit bus from Newark to Manhattan is $2.95. From Hoboken to Manhattan, the bus fare is $1.70. From Newark Airport to Manhattan, it is $7.00. The fare from Newark to Manhattan by Greyhound bus is $7.00.

Several commuter trains are run by the Metro-North Commuter Railroad. Fares for comparable distances on those lines are as follows:

From Mount Vernon, Pelham and New Rochelle--

$4.50, off-peak $3.50.

From Woodlawn and Wakefield--

$4.00, off-peak $3.00.

From Marble Hill and Riverdale--

$4.00, off-peak $3.00.

Where the Long Island Railroad, the busiest commuter railroad of all, competes with the New York subways, its fares are of necessity competitive. However, once Long Island leaves the subway area, its fares escalate rapidly. For example, the fares from Valley Stream, Garden City and Mineola are $5.00.

Although the territories covered by the several lines of New Jersey Transit that connect with PATH at the Penn Station in Newark are not the same as that covered by PATH, a brief look at their fares for distances comparable to the distance between Newark and New York is enlightening. Thus, on the Raritan Valley Line, the fare from Garwood to Newark is $2.50. On the North Jersey Coast Line, the fare from Avenel to Newark is $3.00. On the Northeast Corridor Line, the fare from Rathway to Newark is $2.50. Several other New Jersey Transit Lines connect with PATH at Hoboken. On the Boonton Line, the fare from Glen Ridge to Hoboken is $2.50. On the Pascack Valley Line, the fare from Teterboro to Hoboken is $2.50.

Why is it that among all these commuter lines, PATH's fare of $1.00 stands alone? The answer is clear; the motorists and truckers using the bridges and tunnels are being forced to pay the difference. I cannot accept this as just and reasonable.

I believe that our review of the district court's decision should be no less demanding than that court's review of PATH's decision. We should not accept the district court's generalized conclusion that PATH's $1.00 fare is reasonable without being satisfied that it is supported by adequate underlying facts. Schneiderman v. United States, 320 U.S. 118, 129-31, 63 S. Ct. 1333, 4338-40, 87 L. Ed. 1796 (1943); Russo v. Central School Dist., 469 F.2d 623, 628-30 (2d Cir.1972); Monarch Beverage Co. v. Tyfield Importers, Inc., 823 F.2d 1187, 1192 (7th Cir.1987). Unlike my colleagues, I find such support totally lacking in the instant case.

The ultimate irony of my colleagues' opinion is its recital of the benefits allegedly flowing to motorists and truckers proceeding across the George Washington and Staten Island Bridges as a result of the $1.00 PATH tolls and the concomitant $3.00 bridge tolls. Over one and one-half million users of these bridges are protesting these tolls, and they do not feel benefited. Indeed, their sentiments would find eloquent expression in the ancient proverb that a man can handle his enemies if someone will only protect him from his friends.

Although it plays no part in my decision, I cannot close this dissent without a reference to recently published reports that the Port Authority has adopted the practice of closing some of the toll booths on the George Washington Bridge because of insufficient budgeted funds. To the longsuffering users of the Bridge whose money is being used to subsidize PATH, this must be like having salt rubbed in their wounds.

I would hold, in accordance with our prior decision, that, until Congress authorizes the Port Authority to do otherwise, the Authority must limit the rate base for the George Washington and Staten Island Bridge tolls to the bridges, the tunnels and the bus terminals.

Alternatively, I would remand to the district court with instructions to return the matter to the Port Authority for reconsideration of all the tolls and the making of a reasonable and just allocation of fares and tolls, supported by complete and adequate factual findings.