UNITED STATES COURT OF APPEALS FOR THE D.C. CIRCUIT


EXXON CO UNITED STATES

v.

FERC


95-1520c

D.C. Cir. 1999


*	*	*


Sentelle, Circuit Judge: Exxon Company, U.S.A. and  Tesoro Alaska
Petroleum Company petition for review of the  Federal Energy
Regulatory Commission's ("FERC" or  "Commission") order revising the
valuation methodology for  specified grades of petroleum products
after our partial re- mand of the Commission's earlier order adopting
the distilla- tion method for determining compensation due shippers on
 the Trans Alaska Pipeline System for differences between the 


oil streams injected and oil streams received. See Order  Modifying and
Adopting Contested Settlement Proposal,  Trans Alaska Pipeline Sys.,
65 FERC p 61,277 (1993) ("1993  Order"), approved in part and remanded
in part, OXY USA,  Inc. v. FERC, 64 F.3d 679, 684 (D.C. Cir. 1995)
("OXY"). In  the order before us, FERC approved with modifications a 
contested settlement over the objection of petitioners. We  grant the
petition for review in part and vacate and remand  for further
proceedings those parts of FERC's order approv- ing the use of proxies
for the market valuation of one grade of  petroleum product and the
decision to apply the settlement  prospectively only.


I. BACKGROUND


The Trans Alaska Pipeline System ("TAPS") provides the  only
commercially-viable method for moving crude oil pumped  from the oil
fields on Alaska's North Slope to the shipment  point at Valdez,
Alaska, the Alaskan gateway to the world  market. Several oil
companies own interests in various oil  fields on the North Slope. The
oil in those fields differs  significantly in quality, but the
realities of shipping that oil on  the single pipe of the TAPS
requires the blending of the oil  streams from different fields.
Unlike packages shipped by a  common carrier, the oil streams cannot
be segregated during  shipping, and the blended streams cannot be
separated at the  Valdez end of the pipeline. Instead, at the Valdez
end of the  pipeline, each shipper receives a quantity of the blended 
common stream equivalent to the amount it injected at the  North Slope
end. Companies that inject higher quality crude  receive oil at the
Valdez end of the pipeline identical in  quality to that received by
companies that inject lower quality  crude oil. The TAPS carriers file
tariffs specifying how the shippers will compensate each other for
these differences in quality,  and their methodology must be approved
by the Commission  pursuant to its authority under the Interstate
Commerce Act  ("ICA"), 49 U.S.C.app. s 1 et seq. See also Department
of  Energy Organization Act, Pub. L. No. 95-91, s 402(b), 91  Stat.
565, 584 (1977), codified at 42 U.S.C. s 7172(b) (1988)  (repealed


(transferring authority to regulate oil pipeline rates under the  ICA
from the Interstate Commerce Commission to FERC);  Exxon Pipeline Co.
v. United States, 725 F.2d 1467, 1468 n.1  (D.C. Cir. 1984)
(explaining transfer of authority). TAPS has  created a system which
requires companies injecting lower- quality oil to compensate
companies injecting higher-quality  oil by creating a "Quality Bank,"
which awards shippers  credits for high-quality oil and debits for
low-quality oil. The  TAPS Quality Bank is an arrangement that "makes
monetary  adjustments [among] shippers in an attempt to place each in 
the same economic position it would enjoy if it received the  same
petroleum at Valdez that it delivered to TAPS on the  North Slope."
OXY, 64 F.3d at 684. While this is simple  enough in concept,
determining the relative value of the  injected streams is in fact a
complex technical task. There is  no independent market to set the
relative price of the various  streams of North Slope crude because
the crude is not sold  until after it is commingled and brought to
Valdez. When the  system was originally created, the relative value of
oil was  determined by the "API gravity"1 of the oil because lighter, 
high-gravity crude is generally more valuable than heavier, 
low-gravity crude. See id. at 685. The "straight-line gravity  method"
measured the gravity of each incoming stream and  compared it to the
gravity of the oil received by that shipper  at the far end, and
determined Quality Bank credits or debits  accordingly. See id. In
1989, however, OXY USA and  Conoco, Inc. challenged this methodology,
and in 1991 a  FERC Administrative Law Judge ("ALJ") determined that
it  "no longer yield[ed] a just and reasonable result." 57 FERC  p
63,010, at 65,049-50, 65,052-53 (1991). (For a full explica- tion of


The majority of North Slope shippers in an attempt to  settle the
tariff dispute proposed abandoning the straight-line 




__________

n 1 API gravity is a measure of density created by the American 
Petroleum Institute. Under API gravity analysis, unlike the more 
familiar concept of specific gravity, a higher number indicates a less
 dense crude oil or petroleum product.


gravity method in favor of a "distillation" or "assay" method- ology,
which would value crude oil based on the market price  of the various
component products (called "cuts") created  when the crude oil is
heated to a series of specific tempera- tures and the evaporated
products produced at each tempera- ture are recondensed. See OXY, 64
F.3d at 687. The five  cuts created by this process at the lower
boiling points-- propane, isobutane, normal butane, natural gasoline,
and  naphtha--and one of the heavier cuts, gas oil, are not at issue 
here, as we upheld the method of valuing those cuts in our  earlier
review. See id. at 701. We vacated and remanded for  further
proceedings as to distillate and residual fuel oil ("re- sid").


A. Distillate


Under the original 1993 settlement offer, the distillate cut  included
the portion of the stream that evaporated between  350 and 650 degrees
Fahrenheit. Under the 1993 settlement  order, FERC split this proposed
cut into two cuts, light  distillate (350-450 degrees) and heavy
distillate (450-650 de- grees). FERC determined that it would price
light distillate  as jet fuel and heavy distillate as No. 2 fuel oil,
the products  into which those cuts are normally refined, without
adjust- ment for processing costs. See 1993 Order, 65 FERC  p 61,277,
at 62,288. We rejected that methodology because  each cut would
require further processing to reach the quality  required for the
proxy product. See OXY, 64 F.3d at 693.  Because the settlement as
modified by FERC essentially  valued a raw material as if it were a
finished product, we  determined that it overvalued these heavier
cuts, resulting in  a windfall to those shippers whose streams
contained the  highest relative proportion of heavy crude. See id.
Although  we recognized that we could not require FERC to achieve a 
perfect method of valuing petroleum streams, particularly  streams
including cuts without a market, we nonetheless held  that FERC must
be consistent in its methodological choices.  That is, if the
Commission chose to value a portion of the cuts  at market without
adjusting for processing costs, then it  must, at least "to the extent
possible," attempt to approxi- mate the market value of other cuts


at 694. That is, the Commission cannot "consistent with the 
requirement of reasoned decisionmaking, value some cuts  precisely and
others haphazardly." Id. We therefore re- manded the distillate
valuation for further consideration by  FERC.


B. Resid


As the name implies, the residual, or "resid," cut consists of  the
portion of the petroleum stream remaining after distilla- tion of all
other cuts at lower boiling points. In the 1993  settlement order,
FERC split the resid into two cuts--light  resid (1,000 to 1,050
degrees Fahrenheit) and heavy resid (all  remaining material). The
order valued these cuts in relation  to the market price of proxies:
No. 6 fuel oil for light resid  and FO-380 for heavy resid with no
adjustment for the  processing necessary to receive these market
prices. We  upheld FERC's decision to create a separate light resid
cut,  but vacated the valuation of that cut at the price of No. 6 fuel
 oil as we found that the record did not disclose a relationship 
between the price of that purported proxy and the value of  the cut.
Likewise, we concluded that the record did not  demonstrate that
FO-380 was a reasonable proxy for heavy  resid because the market
price of FO-380 bore only a limited  and unquantified relation to the
value of heavy resid as a  blending component. See id. at 695. While
we concluded  that expert testimony in the record supported a
"conclusion  that FO-380 and the 1050+ resid share some physical prop-
erties," it did not even suggest that "the two materials have  equal
or even near-equal market values." Id. We therefore  remanded the
valuation of the resid cuts to the agency for  further proceedings


In our review of FERC's order approving the 1993 settle- ment, we
rejected not only the specifics of the FO-380  comparison, but also
FERC's decision to value resid based on  its use as a feedstock for
"cokers," refinery equipment which  breaks resid down even further
into lighter fuel products and  a heavy residue, which might be
asphalt at some plants, or  other materials with differing uses. Exxon
and others ar- gued that resid should be priced at its marginal use


which Exxon claimed was as a blending component for  FO-380. When
remanding, we observed that this economic  argument, while it might
not by itself carry the day, did  possess enough "analytical force"
that the Commission should  on remand "explicitly address whether the
marginal use of  1050+ resid should be taken into account in that
cut's  valuation methodology." Id.


C. FERC's Proceedings on Remand


In response to our opinion, FERC initiated settlement  proceedings
regarding these remanded issues. When this  effort failed, FERC set
the matter for hearing. At the same  time, the Commission's Chief ALJ
made further attempts to  secure a settlement. The parties filed three
separate settle- ment proposals, one by nine parties2 ("the Nine Party
Settle- ment"), and unilateral proposals from Exxon and Tesoro.  The
ALJ provided opportunity for all parties to file materials  in support
of or in opposition to the settlement offers. Fol- lowing the
submissions, the ALJ heard oral argument and the  parties filed
supplemental briefs. See Certification of Con- tested Settlement and
Ruling on Motion to Omit the Initial  Decision, Trans Alaska Pipeline
Sys., 80 FERC p 63,015, at  65,212-13 ("1997 Opinion").


The ALJ ultimately certified the Nine Party Settlement to  the
Commission, and opted not to certify the unilateral pro- posals from
Exxon and Tesoro, finding that legal precedent  required this decision
and that in any event the proposals  were biased in favor of the
proposing parties. The ALJ  reviewed the record in detail and
determined that the only  issues properly before him were the remands
for valuation of  light and heavy distillate and light and heavy
resid. He  found that the Nine Party Settlement's proposed valuations,
 which follow, were fair and reasonable and supported by 




__________

n 2 The nine settling parties are Amoco Production Company,  ARCO
Alaska, Inc., BP Exploration (Alaska), Inc., MAPCO Alaska  Petroleum,
Inc., OXY USA, Inc., Petro Star, Inc., Phillips Petrole- um Company,
the State of Alaska, and Union Oil Company of  California. See 1997
Order, 81 FERC p 61,319, at 62,458 n.5.


record evidence. See 1997 Opinion, 80 FERC p 63,015, at  65,233.


Light distillate: valued based on a weighted average of  the West Coast
and Gulf Coast prices of jet fuel, adjust- ed by 0.5 cents per gallon
to reflect processing costs.


Heavy distillate: valued based on weighted average of  the West Coast
price of Waterborne Gasoil, reduced by 1  cent per gallon to reflect
processing costs and the Gulf  Coast price of No. 2 fuel oil reduced
by 2 cents per gallon  to reflect processing costs. (The processing
costs were  based on the testimony of Nine Party expert witness  John
O'Brien who stated that ANS crude oil needed to  be processed to reach
the 0.5 percent level for sulfur  demanded by the market. )


Light resid (1000 degrees F to 1050 degrees F): The  1993 settlement
had eliminated separate treatment of  light resid and combined it with
the 1050+ cut. The  Nine Party Settlement approved by the ALJ instead 
rolled it into the Vacuum Gas Oil ("VGO") cut, by raising  the top end
of that cut to 1050 degrees, which the nine  parties claim conforms
with industry practice.


Heavy resid (1050+): continued use of the West Coast  price of FO-380
as a West Coast reference price, sub- tracting 4.5 cents per gallon as
a processing cost. Added  Gulf Coast 3 percent sulfur No. 6 fuel oil
as a Gulf Coast  reference product, and adjusted that figure by the
same  4.5 cents.


The ALJ noted that the nine parties supported the settle- ment only if
it applied prospectively. See id. at 65,241. The  ALJ determined that
the remand did not require that the  new methodology be applied
retroactively and that the Com- mission retained the discretion to
determine when to make  the settlement effective. See id. at 65,243.
The ALJ also  recommended prospective application under the circum-
stances. See id.


The Commission reviewed and accepted the ALJ's recom- mendations as to
each valuation, finding in its order that each 


determination was based on substantial evidence. FERC  found that there
was no active market for resid, and opted to  price resid based on its
value as a coker feedstock. FERC  determined that the two reference
products were the actively- traded petroleum products that had
physical characteristics  most resembling resid, and used these
adjusted prices as a  proxy for the value of resid as a coker
feedstock. It also  decided to apply the new rates prospectively,
stating that this  was consistent with the 1993 Order applying the new
rates  prospectively, which was affirmed by this court in OXY.  "[The
new settlement] does not change the methodology to be  used, but
modifies how to value the remanded cuts." See  1997 Order, 81 FERC p
61,319, at 62,467. The Commission  noted that the TAPS Quality Bank
was sui generis, so  precedents cited by Exxon and Tesoro as
supporting retroac- tive application of the new methodology were not


II. STANDARD OF REVIEW


The standard of review applicable to FERC's approval of  this proposed
settlement of the issues remaining on remand is  the same as it was in
OXY. FERC's decision to approve a  portion of a contested settlement
must be supported by  substantial evidence, and we must set aside
FERC's approval  if it was "arbitrary, capricious, an abuse of
discretion, or  otherwise not in accordance with law." 5 U.S.C. s
706(2)(A),  (E). Our inquiry under the arbitrary and capricious test
is  "narrow and a court is not to substitute its judgment for that  of
the agency." Motor Vehicle Mfrs. Ass'n of the United  States, Inc. v.
State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43  (1983). Where, as in
the instant case, the analysis to be  performed "requires a high level
of technical expertise, we  must defer to the informed discretion of
the responsible  federal agencies." Marsh v. Oregon Natural Resources 
Council, 490 U.S. 360, 377 (1989) (internal quotation marks  omitted).
Nonetheless, the Commission must engage in ra- tional decisionmaking,
see, e.g., State Farm, 463 U.S. at 43;  OXY, 64 F.3d at 690. We held
in OXY that the agency had  supplied a reasoned analysis for changing
its prior policies  when it adopted the distillation methodology. See


F.3d at 690. However, more important for purposes of the  petitions now
before us, we granted the petitions for review of  the 1993 Order to
the extent that they challenged the Com- mission's methods of valuing
the distillate and resid cuts.


III. CHALLENGES TO THE NEW SETTLEMENT


The petitioners make multiple arguments challenging the  valuation of
specific cuts and FERC's failure to require that  qualitative
differences between the same cuts of different  streams be considered
when determining the relative value of  each stream. They argue that
FERC acted arbitrarily by  failing to value resid based on its
marginal use as a fuel oil  blendstock instead of as a coker
feedstock; improperly failed  to account for differences in quality
among the the same cuts  of different streams when valuing resid;
improperly chose a  price proxy for its value as a coker feedstock;
and failed to  address challenges to the methodology for determining
resid's  value as a coker feedstock. The petitioners also challenge 
FERC's decision to implement the new valuation methodolo- gy
prospectively only. We address first the valuation chal- lenges, and
uphold the agency's decisions as supported by  substantial evidence
with the exception of the use of FO-380  less 4.5 cents and 3 percent
sulfur No. 6 fuel oil less 4.5 cents  as proxy prices for heavy resid.
The adjusted valuation  solves none of the problems we identified in
our prior opinion  because there is no evidence that the prices of the
reference  products, even after the 4.5 cents adjustment, bear any 
rational relationship to the market value of resid. We there- fore
vacate and remand the portion of FERC's order affect- ing the


IV. INTRA-CUT QUALITY DIFFERENCES


Exxon3 argues that FERC's failure to account for differ- ences in
quality among the heavy distillate cuts of the individ- ual streams
before they are commingled in the TAPS com-




__________

n 3 Exxon and Tesoro filed a joint petition for review. For simplici-
ty's sake, we will refer to the joint arguments of the two petitioners
 as Exxon's arguments.


mon stream violates the terms of our earlier remand in OXY,  and is
arbitrary and capricious. We disagree.


Petitioners claim that the goal of the Quality Bank is to  place an
accurate value on the streams flowing into the TAPS,  and failure to
account for quality differences in the distillate  cuts of the streams
coming from different oilfields is not  reasoned decisionmaking. We
disagree that Exxon's argu- ment follows logically from our remand. In
OXY, we recalled  that the goal of the Quality Bank is "to assign
accurate  relative values," 64 F.3d 693 (emphasis added), to the
diverse  streams delivered to the pipeline. We vacated in part the
last  order because the methodology approved therein had favored  one
class of cuts above others. We remanded in order that  FERC might
provide a methodology with a reasoned relative  uniformity, knowing
that absolute precision at any level of the  cuts was unachievable.
That is, we did not remand because  the old method was inaccurate, but
because it was unfairly  nonuniform. To have demanded 100 percent
accuracy would  have been to hold the agency to "an impossibly high
stan- dard." Id. at 694. The specific purpose in our remand was  to
require the agency to resolve the relative overvaluation of  some
cuts, which were valued at the market price for their  proxy despite
the fact that significant processing was re- quired to bring those
products up to a market standard.  Exxon seeks to expand the duty of
the Commission to refin- ing the degree of distinction among component
streams with- in individual cuts. Specifically, Exxon seeks to have us
 vacate FERC's order insofar as it does not recognize and  adjust for
differences in the sulfur content of distillate as a  key factor in
determining market value. Part of the adjust- ment to the per-barrel
price of distillate is to account for  removing sulfur so that it can
be sold as jet fuel or No. 2 fuel  oil. In implementing that
methodology, FERC assumed that  all streams had the same sulfur
content, when Exxon had  shown that such was not the case. Exxon
argues that FERC  should not use the sulfur content of the commingled
streams  when determining the value of the cut, but must determine 
the sulfur content and thus the value of the distillate cut of  the


Because some streams have a higher sulfur content, they  would require
more processing and consequently have a lower  value once processing
costs were factored into the per-barrel  price. Other streams with a
lower sulfur content would have  a higher value because no further
processing would be needed  to bring the oil up to the quality of the
proxy product.


Exxon further argues that treating all of the streams as if  they have
the same sulfur content violates OXY, which calls  for accurately
valuing the streams; that it is arbitrary and  capricious because it
makes assumptions contrary to fact;  and that FERC's failure to even
consider the issue is arbi- trary and capricious. Specifically, Exxon
argues that FERC  improperly determined that the scope of its actions
was  limited by the terms of our remand, but that in any event,  FERC
cannot claim that it addressed only the issues required  by the court
because it did more than we ordered when it  changed the West Coast
proxy for heavy distillate, even  though no party challenged the one
adopted in the 1993 and  1994 orders, and eliminated the light resid
cut, even though it  was affirmed in OXY. Exxon contends that having
opened  the door, so to speak, FERC was obligated to consider the 
information provided by Exxon and Tesoro about the differ- ences in
quality among the streams because it has an obli- gation under the ICA
" 'to ensure that pipeline rates are just  and reasonable.' " OXY, 64
F.3d at 690 (quoting Texas  Eastern Transmission Corp. v. FERC, 893
F.2d 767, 774 (5th  Cir. 1990)). Exxon argues that refusing to
consider the  quality differences was therefore arbitrary and an abuse
of  discretion. In its 1997 Order, FERC noted that it had  rejected
the same argument in its 1993 Order, and that we  had not reversed or
vacated that ruling. Exxon argues  nonetheless that by adjusting the
market prices of the proxies  to account for removing sulfur, FERC
itself has now deter- mined that sulfur content is an important aspect
of valuing  heavy distillate.


We reject Exxon's argument that FERC's failure to differ- entiate
between the streams was arbitrary and capricious. In  OXY, we required
FERC to take into account the significant  processing costs that
rendered its unadjusted use of a proxy 


product unreasonable in relation to the valuation of other  portions of
the stream. Exxon's contention that FERC must  value each stream at
the wellhead based on its individual  sulfur content calls for more
than we required. We did not  hold in OXY that differences in quality
between the streams  must be considered, and do not do so now.
Inherent in our  approval of FERC's adoption of the distillation
methodology  in OXY was our approval of the agency's conclusion that
there  was no need to consider intra-cut quality differences, and that
 the agency properly determined that the relative proportions  of the
cuts in each stream is sufficiently accurate as a method  of
determining the relative value of the streams. See 65  FERC p 61,277,
at 62,287 (1993), and 66 FERC p 61,188, at  61,240 (1994). In any
event, it was not arbitrary and capri- cious to determine the value of
each cut in the TAPS stream  after it has been mixed, instead of
separately valuing the cuts  of each stream. The fact that a more
precise method exists  for determining the relative value of the
streams does not  render the decision to adopt a less accurate, but
more admin- istrable, method arbitrary and capricious. FERC has opted 
to use a magnifying glass to determine the values of the  streams, and
we will not fault it for not using a microscope.


We also uphold against challenge FERC's two changes to  the price of
heavy distillate, both of which are supported by  the record. FERC
changed the reference price for the West  Coast from No. 2 fuel oil to
Waterborne Gasoil, and adjusted  the price of Waterborne Gasoil by one
cent per gallon and the  Gulf Coast price of No. 2 fuel oil by two
cents to account for  processing. See 1997 Order, 81 FERC p 61,319, at
62,460.  These adjustments were based on the testimony of expert 
witnesses John O'Brien and Christopher Ross. Ross testified  that
these products most closely resembled Alaskan North  Slope ("ANS")
heavy distillate, see Affidavit of Christopher E.  Ross, p 19 (Jan.
29, 1997), and O'Brien testified that the ANS  heavy distillate cut
required treatment to reach the necessary  sulfur level, see Affidavit
of John O'Brien, WW 13-15 (Jan. 28,  1997). These decisions were
supported by adequate record  evidence and we uphold the agency.


V. RESID CUT VALUATION ISSUES


A. Exxon and Tesoro's Challenges


In OXY, we noted that resid like distillate did not trade on  an open
market and therefore was difficult to evaluate.  Nonetheless, and even
in the face of "the deference we owe  the Commission's judgments," we
concluded that the 1993  settlement approach to valuation of resid did
not "satisfy the  APA's basic requirement of reasoned decisionmaking."
OXY,  64 F.3d at 694 (citing State Farm, 463 U.S. at 43). We 
therefore remanded that portion of the assay methodology to  the
Commission for further consideration.


The method before us in the present review fares no better  than the
last, and for the same reasons: even with the 4.5  cents per gallon
adjustment, "the record demonstrates no  more than that the price[s]
of FO-380 [or No. 6 fuel oil]  bear[ ] some remote relationship to the
value of 1050+ resid  as a feedstock." Id. at 695. We remand FERC's
decision to  value resid at the price of FO-380 less 4.5 cents on the
West  Coast and Waterborne 3% sulfur No. 6 fuel oil less 4.5 cents  on
the Gulf Coast. The figures derived from the use of these  proxies
with a subsequent adjustment do not bear a demon- strated relationship
to the value of resid, either as a coker  feedstock or as a blending
agent for fuel oil. Exxon and  Tesoro raise multiple challenges to
FERC's valuation process  for this cut.


1. Marginal Use


Exxon argues that FERC erred again, as it did in the 1993  Order in not
employing the marginal use of resid as a  blending agent for fuel oil
rather than its value as coker  feedstock in establishing the
valuation methodology for that  cut. Exxon contends that the error is
a fundamental one in  that the ALJ's finding, adopted by the
Commission, that  there is no active market for resid is flawed. In
Exxon's  view, although there are few trades of resid, there is in
fact a  market, and a sparsity of open trades is only due to the fact 
that the refiners who use resid rarely need to purchase it  from
others because they already obtain it as a byproduct of 


their own refining operation. Exxon further argues that  there are
formulae that can be used to derive resid's value as  a blendstock
despite the absence of market trades. Thus  Exxon prays the court to
vacate the relevant portion of  FERC's order and remand the
controversy for valuing of  resid as a blendstock.


FERC responds that there was conflicting evidence regard- ing the
existence of a market for ANS resid, and the ALJ and  the Commission
reasonably adopted the testimony of Nine  Party witnesses A.L.
Gualtieri and Benjamin Klein, who  testified that resid was rarely
traded, and was instead used as  a coker feedstock. See 1997 Opinion,
80 FERC p 63,015, at  65,238-41. The ALJ also determined, based on the
record,  that it was inappropriate to value resid based on its
marginal  use as fuel oil blendstock because most of the refineries
did  not seek to purchase resid but created it as part of their 
refinery process. See id. 65,240. The absence of an active  market for
resid made the economic principle of marginal use,  which depends on a
liquid market, unreasonable in this  circumstance. See id.


We see no reason to disturb FERC's adoption of the ALJ's  determination
that resid is best valued based on the market  value of its
constituent products. The expert testimony of  Klein constitutes
substantial evidence in support of FERC's  decision that marginal use
analysis does not require the  valuation of resid as a blendstock.


2. Conradson Carbon Residue Content


As with distillate, Exxon argues that FERC arbitrarily  ignored quality
differences in the streams which affect the  value of the different
cuts. The Conradson Carbon Residue  Content ("CCR") of resid affects
its value, and the different  streams delivered to the TAPS
undisputedly have differing  CCR content. Exxon reiterates the
argument it made con- cerning sulfur that failing to account for
differing CCR  content was arbitrary and capricious. The CCR content 
figure used by FERC was not even derived from the oil  shipped over
TAPS, but from a blend used by an expert  which included other crude
oils. FERC responds that it 


properly rejected the suggested intra-cut differentials based  on CCR
content for the same reasons it rejected the quality  differentials
based on sulfur content. For the reasons stated  in Parts III and IV
above, we hold that FERC was not  required to consider intra-cut
differences in CCR content  when determining market value.


3. Choice of Proxies


Exxon next argues that FERC acted arbitrarily when it  chose to use the
adjusted price of FO-380 as a proxy for  valuing resid as a coker
feedstock. In OXY, we found that  using the unadjusted market price of
FO-380 as a proxy was  arbitrary and capricious. The 4.5 cents
adjustment now  adopted is arbitrary for the same reasons. There is no
 demonstrated relationship between the value of FO-380 and  coker
feedstock other than an observed rough correlation in  price, and even
the data relied on by FERC shows inconsis- tent relationships in the
price of FO-380 and the coker  feedstock values calculated by the
experts. Exxon argues  that determining resid's value as a coker
feedstock "requires  determining the identity, quantity, and value of
products  produced in a coker from resid and subtracting from the 
value the costs of producing those products and placing them  in a
marketable condition." See Joint Brief of Petitioners  Exxon Company,
U.S.A. and Tesoro Alaska Petroleum Com- pany at 42. Exxon also argues
that FERC chose the wrong  feedstock to value because it used a blend
of crudes which  would be used by a hypothetical refinery, rather than
actual  individual North Slope crude streams. Exxon further con- tends
that it presented numerous challenges to the methodol- ogy ultimately
adopted by FERC, showing inaccuracies in the  expert's assumptions
regarding cost calculations, product out- puts and product yields.
Finally, it argues that because the  ALJ never allowed discovery, it
could not replicate the ex- pert's computer modeling on the PIMS
system (a standard- ized petroleum industry modeling system used to
calculate  refinery needs and outputs). The ALJ and the Commission 
did not specifically address these arguments, which Exxon  contends
makes their decisions arbitrary and capricious.


FERC responds that the 4.5 cents per gallon adjustment to  the price of
FO-380 on the West Coast and No. 6 fuel oil on  the Gulf Coast as
proxies for resid was reasonable, based on  expert witness O'Brien's
testimony and administrative ease.  These are the lowest-quality
products actively traded, and the  adjustment was within the range of
variation between the  calculated value of resid as a coker feedstock
and the per- gallon price of FO-380. See Ross Affidavit p 21. O'Brien 
derived the calculated value of resid as a coker feedstock  using the
PIMS model and compared those calculated values  to the market price
of FO-380 over the same five-year period.  The relationship varied
from resid being worth $1.21 per  barrel more than FO-380 in 1993 to
being worth $3.01 per  barrel less than FO-380 in 1995, and averaged
being worth  $1.12 per barrel less over the five-year period. See 1997
 Opinion, 80 FERC p 63,015, at p 65,239 (citing O'Brien Affida- vit WW
56-598 Exhibit QB ar-23). O'Brien testified that the  4.5 cents per
gallon adjustment (equal to $1.89 per barrel less  than FO-380)
proposed by the Nine Party Settlement fell  within the observed range
of variation over the five-year  period and was therefore reasonable.
See id. FERC also  notes that Exxon and Tesoro both suggested a method
that  tied the price of heavy resid to FO-380. The difference is  that
Exxon uses a complex formula to adjust the price.4


B. Analysis


While we find substantial record evidence supporting the  intermediate
steps FERC took in determining the value of  resid--i.e., its
determinations that no active market exists,  that resid is best
valued as a coker feedstock rather than as a  blender for fuel oil,
and that FO-380 and No. 6 fuel oil are the  actively-traded products
in the relevant markets most similar  in physical characteristics to
resid--we cannot conclude that  the last step follows logically from
these premises. We 




__________

n 4 FERC's suggestion that Tesoro and Exxon somehow validated  their
choice of FO-380 as a reference product is misleading because  Exxon
and Tesoro's use of FO-380 as a reference price ties the  value of
resid to the value of FO-380 when valuing resid as a  blendstock for
fuel oil, not as a coker feedstock.


therefore cannot uphold the use of FO-380 less 4.5 cents on  the West
Coast and Waterborne 3% sulfur No. 6 fuel oil less  4.5 cents on the
Gulf Coast as a proxy price for resid.


The 4.5 cents adjustment, while it falls within the range of  the
observed variation, does no more than that. There is no  evidence that
the prices of the proxy products are more than  coincidentally related
to the value of resid as a coker feed- stock. Moreover, the calculated
value of resid using the  PIMS model does not even vary consistently
with the price of  FO-380. As petitioners noted when this case was
before us  in OXY, by the same logic we could use the price of coal
with  an adjustment as a proxy for the price of diamonds because  both
are a source of carbon, even if the prices fluctuate  inconsistently.
With only five years' data to consider, the  sample is too small to
convince us that there is some other,  unstated relationship at work
which guarantees that the price  of FO-380 and the value of resid will
correlate consistently  within some specified range. We recognize that
the agency is  addressing the Quality Bank Administrator's concerns
that  more complex systems may give the appearance that the  price of
resid is open to manipulation, and thus is seeking a  product that is
traded on the market to use as a proxy, which  would allow the Quality
Bank Administrator to perform a  simple market-based calculation when
determining the value  of resid. These goals of administrative
efficiency and objec- tivity do not free the agency from the
requirement that the  chosen proxy bear a rational relationship to the
actual market  value of resid. We remand once again to the agency to 
determine a logical method for deriving a value for resid.  Because we
remand, we do not reach the technical objections  Exxon and Tesoro
raise regarding specific calculations.


VI. TESORO'S INDEPENDENT CHALLENGES


A. Tesoro's Standing


In addition to the arguments raised jointly with Exxon,  Tesoro raises
numerous additional challenges to FERC's  decision. However, before we
address the arguments raised  by Tesoro in its individual brief, we
must consider as a 


threshold matter whether Tesoro has standing to petition us  for
review. Intervenors argue that Tesoro lacks standing  because it is no
longer a shipper on the TAPS system and  therefore no longer has a
legally cognizable stake in the  outcome. As a result, they argue, the
case is moot as to it  and issues raised only by Tesoro are not
properly before us.  Intervenors also argue that because Tesoro passed
its Quality  Bank costs through to its shippers, it was not aggrieved
by  the orders under review.


Tesoro counters that it has standing as a competitor of  MAPCO, one of
the shippers on the TAPS system, which is  subsidized by TAPS because
its stream is overvalued. We  have held that even non-shippers and
competitors may be  within the ICA's zone of interest. See OXY, 64
F.3d at 697.  Tesoro also notes that it currently purchases ANS crude
from  one supplier and hopes to acquire more from another. Teso- ro
Reply Brief at 19 n.10.


The Intervenors are correct that only "aggrieved" parties  may seek
judicial review of a final FERC order issued under  the ICA. See 28
U.S.C. s 2344; OXY, 64 F.3d at 696; Shell  Oil Co. v. FERC, 47 F.3d
1186, 1200 (D.C. Cir. 1995). We use  traditional standing principles
to determine if a party is  indeed aggrieved. See OXY, 64 F.3d at 696;
Water Transp.  Ass'n v. ICC, 819 F.2d 1189, 1193 (D.C. Cir. 1987). To
be  aggrieved, Tesoro must have suffered an "injury in fact" 
traceable to FERC's action, a decision in its favor must be  capable
of redressing that injury, and its interest must be  within the zone
of interests protected by the statute. Tesoro  has shown that it would
suffer competitive injury if other  shippers were advantaged by unfair
Quality Bank valuations,  a decision on our part altering those
valuations would redress  that injury, and the ICA permits a very
broad range of  parties to complain to FERC about pipeline operations.
The  ICA permits the Commission to respond to complaints about 
"anything done or omitted to be done by any common carri- er" subject
to the statute lodged by, inter alia, "[a]ny person,  firm,
corporation, company, or association." 49 U.S.C.app.  s 13(1). Tesoro
has standing to challenge the decision here.


B. Tesoro's Position


Tesoro marshals additional attacks on FERC's approval of  the
settlement, some technical and some that are arguably  procedural.


1. Considering Processing Costs for Only Two Cuts


Tesoro argues that FERC erred in singling out the light  and heavy
distillate cuts for processing cost calculations when  processing
costs associated with other cuts are ignored. It  argues that this
violates the requirement in OXY that streams  be valued equally. In
OXY we remanded the light distillate  and heavy cuts for new valuation
because further processing  was required before they could be sold as
jet fuel and No. 2  fuel oil respectively. Tesoro now claims that FERC
arbi- trarily ignored the question of whether further processing  was
needed before the other cuts could be sold as the proxy  products FERC
used to value them. Failing to do so, it  claims, skews the valuation
in favor of the heavier streams.  This argument fails to comprehend
our earlier opinion.  There we upheld the agency's finding that the
lighter cuts  were of sufficiently comparable quality to the market
proxies  that no further processing was needed, and therefore no cost 
adjustment was needed. Essentially, the market price was  correct
because in those instances the distillation method  resulted in a
market-ready product. We will not reexamine  this issue now. For the
reasons given above in Parts III, IV,  and V.A.2, we do not entertain
the argument that quality  differences between the streams must be


2. Costs of Sulfur Removal


Tesoro argues that internal inconsistencies in the Nine  Party data
show that the processing costs for sulfur removal  are not credible,
specifically because there is a higher per- unit cost to remove sulfur
from heavy distillate than from  resid. Tesoro presented evidence
challenging these calcula- tions, which the ALJ and FERC failed to
fully address.


FERC responds that Tesoro's argument that there are  inconsistencies in
O'Brien's cost calculations for sulfur remov-


al was never raised before the Commission, and cannot be  raised now
before the court. If the issue was preserved, the  agency argues that
Tesoro has produced no evidence showing  that the calculations are
incorrect, and that the agency could  reasonably have adopted
O'Brien's calculations.


We hold that Tesoro preserved this issue for review when it  argued
before the Commission that there was "no way,  absent discovery, to
determine that O'Brien's cost estimates  are not totally arbitrary"
and that the conflicting testimony of  its experts supported a lower
cost per unit for removing  sulfur. Motion of Tesoro Alaska Petroleum
Company for  Expedited Reconsideration and Remand or to Permit Appeal 
Concerning Certification of Nine Party Settlement WW 36-37  (Oct. 15,
1997). As for the merits of the issue, we hold that  FERC reasonably
relied on the testimony of Nine Party  witness O'Brien in reaching the
adjustment. Witness O'Brien  testified that different methods would be
needed to bring the  two products into compliance. Heavy distillate
could be  blended with a lighter product to bring it into compliance 
with the 0.5% market tolerances for sulfur in West Coast  Waterborne
Gasoil, the reference product on the West Coast.  However, such
blending would not be economically feasible to  bring it down to the
0.2% sulfur content of Gulf Coast No. 2  fuel oil, the Gulf Coast
reference product, so it would have to  be processed to remove the
excess sulfur. See 1997 Opinion,  80 FERC p 63,015, at 65,234; O'Brien
Affidavit WW 13-15.  This difference in approach accounts for the
difference in  cost. Thus, there is no inconsistency warranting the


3. Processing Costs for Light Distillate


Tesoro argues that FERC arbitrarily and capriciously ac- cepted the
Nine Parties' processing cost adjustment for light  distillate. Tesoro
argues that its expert testified that no  further processing was
required for light distillate to meet  the requirements for jet fuel,
the proxy product used for  valuation of the light distillate cut.
FERC arbitrarily accept- ed the Nine Parties' experts' claims that 0.5
cents per gallon  in processing was required before the cut would meet


standard. Tesoro also argues that its expert pointed out  unreasonable
additions to the cost of the processing, such as  unnecessary pumping
and inflated administrative costs, and  that FERC accepted this flawed
estimate without considering  contrary evidence and thus failed to
satisfy the substantial  evidence standard. We find this objection to
be without  merit. There is substantial record support for the Commis-
sion's determination that a 0.5 cent/gallon adjustment was  required
to account for the processing of light distillate into  jet fuel. That
evidence consisted of expert testimony before  the ALJ by Nine Party
witness O'Brien supporting the  processing costs figures eventually
adopted by the ALJ and  thereafter by the Commission. See Reply
Comments of the  Nine Settling Parties in Support of the Nine Party
Settle- ment at 4-5 (Mar. 17, 1997).


4. Coker Feedstock Value Based on Improper Assump- tions and
Calculations Not in the Record


Tesoro next argues that FERC ignored substantial and  important
criticism of the coker valuation of resid. Under  the adopted method,
resid's coker feedstock value is deemed  to be the value of the
products created less the cost of  processing. Tesoro argues that the
other experts' opinions  were based on the wrong mix of product
yields, that the  PIMS model used is not in the record, and that
Tesoro's  expert could not replicate the results. Tesoro also argues 
that its expert showed that the coker operating costs used by  the
Nine Parties' experts were overstated. Because the  PIMS model is not
in the record, FERC could not make a  rational connection between the
facts and the conclusions  drawn therefrom.


Given that we are remanding the question of valuation of  resid because
FERC has not provided a reasoned explanation  for its determination to
set resid's value as a coker feedstock  and to use FO-380 less 4.5
cents on the West Coast and  Waterborne 3% sulfur No. 6 fuel oil less
4.5 cents on the Gulf  Coast as a proxy price, we need not decide this
detailed  factual question, as the factual record may change on
remand.  FERC will necessarily address these issues when it revalues


resid, and such complex technical questions belong first to the 
informed discretion of the agency. See OXY, 64 F.3d at 691.


5. Eliminating the Fuel Oil Cut


Tesoro argues that FERC improperly eliminated the light  resid cut and
determined that the 1000-1050 degree cut  should be valued as VGO. (We
had previously affirmed  FERC's creation of the light resid cut, but
had remanded for  new valuation.) The Nine Parties had suggested this
change,  and FERC approved it. Tesoro argues that the new cut is 
beyond the capability of many refineries. It suggests that  the ALJ
was confused when he determined that this change  was consistent with
the Commission's treatment of this cut.


FERC reasonably found, in resolving this technical matter,  that the
record evidence supports a determination that "the  standard industry
cut point shown on assays is 1050¯, and  that the published
specifications for VGO permit cut points to  1100¯." See 1997 Order,
81 FERC p 61,319, at 62,464. This  finding, coupled with the testimony
of expert witness O'Brien,  see id. at 65,236-37, provided substantial
evidence supporting  the agency's decision that VGO is a permissible
product on  which to base the valuation of 1000 to 1050 degree


6. The Choice of Waterborne Gasoil


Tesoro argues that FERC arbitrarily and capriciously ap- proved the
Nine Parties' selection of Waterborne Gasoil as  the proxy product for
valuing West Coast heavy distillate.  Tesoro argues that Waterborne
Gasoil is not a West Coast  product, but is a Singapore product
created in Singapore and  is thus subject to Far East refining and
market economics.  This, it argues, is inconsistent with the stated
goal of the  settlement of valuing the product on the coast where it
is  "delivered and used." Waterborne Gasoil, a high-sulfur prod- uct,
cannot be sold on the West Coast. See Tesoro Brief at  19-21.


The agency states that "the reference price used is 'Platt's  U.S. West
Coast spot quote for Waterborne Gas oil less 1  cent per gallon for
processing costs.' ... That quoted Platt  West Coast Waterborne Gas
Oil price represents the value of 


significant Gas oil transactions on the United States West  Coast."
1997 Order, 81 FERC p 61,319, at 62,463-64. Wit- ness Ross stated that
the price for Waterborne Gasoil was a  West Coast price, even if the
product was ultimately exported  to Singapore. See Affidavit of
Christopher E. Ross WW 7-10  (Mar. 17, 1997). Given this record
support, we will not  disturb FERC's determination.


7. Inconsistent Treatment of Heavy Distillate and Resid


Tesoro also argues that the valuation of heavy distillate is 
inconsistent with the valuation of resid. West Coast heavy  distillate
is valued based on its marginal use as the lowest- value product
requiring the least processing (high sulfur  Waterborne Gasoil),
whereas resid is valued based on its  highest-value use as a coker
feedstock. Tesoro argues that  this impermissible inconsistent
treatment overvalues the  heaviest streams. This amounts to a
reiteration of the ques- tion addressed above regarding FERC's
determination that it  is appropriate to value resid as a coker
feedstock in the  absence of a liquid market for the product. We
uphold  FERC's decision for the reasons stated above in Section 


8. Naphtha and Gas Oil


Tesoro argues that FERC should have reevaluated other  cuts,
particularly naphtha and gas oil. Specifically, Tesoro  argues that
FERC failed to value these two cuts based on a  weighted valuation of
the prices on both the West Coast and  Gulf Coast, which violates the
"dual-market principle." See  Brief of Petitioner Tesoro Alaska
Petroleum Company at 22.  None of these products are valued based on
Gulf Coast  prices, which overvalues gas oil and undervalues naphtha, 
thus favoring heavy streams. Whatever the merits of these  arguments
might be, the issues they raise are beyond the  scope of the limited
remand, and therefore not properly  before us.


C. Procedural Questions


Tesoro next argues that FERC arbitrarily and capriciously  failed to
provide for adequate procedures to ensure a reliable 


record. Specifically, Tesoro argues that FERC should have  ordered
discovery and hearings with cross-examination to  resolve contested
issues because of the vastly differing posi- tions of the experts.
Live hearings would have permitted the  ALJ to make credibility
determinations, and cross- examination would have permitted Tesoro to
challenge specif- ic portions of the experts' testimony. For instance,
Tesoro  objects that the PIMS computer model is not in the record, 
and thus the assumptions underlying the coker feedstock  valuations
could not be tested,5 and argues that some of the  Nine Parties
advocated higher payments into the Quality  Bank earlier in the
litigation. Tesoro cites Astroline Com- munications Co. Ltd.,
Partnership v. FCC, 857 F.2d 1556,  1571 (D.C. Cir. 1988); Porter v.
Califano, 592 F.2d 770, 783  (5th Cir. 1979); and Xerox Corp. v.
Genmoora Corp., 888 F.2d  345, 355 (5th Cir. 1989), as establishing
the principle that  review of a contested settlement on the merits


FERC responds that the procedures employed by the ALJ  provided ample
opportunity for the parties to advance all  supporting evidence for
their proposals and to illuminate  defects in the counter proposals.
Specifically, the ALJ per- mitted the parties to file affidavits and
other materials in  support of the proposals; the ALJ heard oral
arguments from  all parties in support of the proposals; the ALJ
further  permitted the parties to file post-argument briefs. FERC 
contends that these opportunities were adequate to fulfill all  due
process requirements and allowed the parties to ade- quately present
their positions to the ALJ and the Commis- sion. We agree.


While it is true that live testimony and cross-examination  can
facilitate a fact-finder's attempts to sort out the truth, we  have
not held that such procedures are necessary in all cases.  In fact, we
have held that "FERC may resolve factual issues  on a written record
unless motive, intent, or credibility are at 




__________

n 5 In light of our remand for reevaluation of heavy resid as a coker 
feedstock, the absence of the PIMS model from the record could in  any
event be no more than harmless error.


issue or there is a dispute over a past event." Union Pac.  Fuels, Inc.
v. FERC, 129 F.3d 153, 164 (D.C. Cir. 1997); see  also Louisiana Ass'n
of Indep. Producers & Royalty Owners  v. FERC, 958 F.2d 1101, 1113
(D.C. Cir. 1992) (party may not  complain that it was deprived of a
fair hearing after receiving  notice of expert testimony on which
opposing party relied, an  opportunity to review it, a chance to
submit briefs criticizing  it and evidence opposing it, and the
opportunity to argue  before the Commission). In this case, there is a
dispute  among experts over the proper method for valuing petroleum 
streams. This type of technical dispute is amenable to reso- lution by
resort to the written record, particularly where  Tesoro had
significant opportunities to submit evidence of its  own and criticize
the evidence submitted by the Nine Parties.  We decline to overturn


VII. PROSPECTIVE APPLICATION  OF THE SETTLEMENT


A. Exxon and Tesoro's Position


Exxon and Tesoro argue that FERC committed legal error  when it decided
that it would implement the settlement order  prospectively only. The
method that we found unreasonable  and remanded has been in effect
since 1993, and the Commis- sion stated when it was adopting the
distillation methodology  that in the event it was reversed and Exxon
suffered econom- ic losses, it could correct any legal errors after
the appeal.  See Order on Rehearing, Trans Alaska Pipeline Sys., 66 
FERC p 61,188, at 61,423 (1994). Now, when it has corrected  the legal
errors identified in OXY, the Commission has opted  to apply the new
rates prospectively only, leaving the parties  without remedy for the
years of unlawful valuations, and  granting the settling parties a


Exxon argues that this circuit's precedents require FERC  to return the
parties to the position they would have occupied  had this legal error
not been made. See Public Utils.  Comm'n of the State of California v.
FERC, 988 F.2d 154, 168  (D.C. Cir. 1993) ("CPUC") (citing cases); see
also, e.g., Pan- handle Eastern Pipe Line Co. v. FERC, 907 F.2d 185,
189 


(D.C. Cir. 1990); Office of Consumers' Counsel, State of Ohio  v. FERC,
826 F.2d 1136, 1139 (D.C. Cir. 1987) (per curiam).  This rule is drawn
from "the logic of the statute itself."  Natural Gas Clearinghouse v.
FERC, 965 F.2d 1066, 1074  (D.C. Cir. 1992).


FERC's reasons for refusing to do so, Exxon argues, are  wrong as a
matter of law. First, the agency agreed with the  ALJ that the cases
cited by Tesoro and Exxon are not  dispositive because, while CPUC and
Panhandle "recognized  that the Commission has the authority in some
circumstances  to issue orders which have retroactive effect, neither
of those  cases required it." 1997 Opinion, 80 FERC p 63,015, at 
65,242. Exxon argues that the language from those cases  explicitly
states that "when the Commission commits legal  error, the proper
remedy is one that puts the parties in the  position they would have
been in had the error not been  made." CPUC, 988 F.2d at 168. This use
of the word "the,"  as opposed to "a," proper remedy suggests FERC
must order  retroactive payment when it commits legal error.


Exxon also argues that FERC improperly attempts to rely  on the filed
rate doctrine as mandating prospective applica- tion of its order. See
1997 Order, 81 FERC p 61,319, at  62,467. Exxon argues that despite
its protestations, FERC  has the authority to correct its error, and
that the shippers  had notice that there might be a later correction
to the rate,  which " 'changes what would be purely retroactive
ratemak- ing into a functionally prospective process by placing the 
relevant audience on notice at the outset that the rates being 
promulgated are provisional only and subject to later revi- sion.' "
Natural Gas Clearinghouse, 965 F.2d at 1075 (quot- ing Columbia Gas
Transmission Corp. v. FERC, 895 F.2d  791, 797 (D.C. Cir. 1990)).


Exxon next argues that even if FERC did have discretion  to determine
whether to apply the corrected valuation retro- actively, its failure
to do so in this case amounts to an abuse  of that discretion. FERC
stated as reasons for its decision  the observations that the change
here was one of valuation,  not of methodology, and that the Quality
Bank was sui 


generis. Neither of these reasons, it contends, supports the  decision
not to remedy the injury to Exxon and Tesoro.  Exxon notes that FERC
had retroactively applied adjust- ments in vacuum gas oil rates that
were set under the  distillation method, rendering both justifications
meaningless.  Exxon also points out that FERC does not explain how the
 "sui generis" nature of the Quality Bank has any bearing on  whether
the aggrieved parties should be made whole.


Exxon further argues that the refusal to make the ag- grieved parties
whole violates the central purpose of the  Quality Bank, which was
created as part of FERC's " 'con- tinuing obligation to ensure that
pipeline rates are just and  reasonable.' " OXY, 64 F.3d at 690
(citing 49 U.S.C. s 1(5)  and quoting Texas Eastern Transmission
Corp., 893 F.2d at  774). Moreover, it contends, this abuse of
discretion is  compounded because FERC refused the injured parties a 
stay pending appeal in 1994 on the basis that it could correct  any
legal errors later found on appeal.


Exxon cites a string of our precedents holding that it is  proper to
correct such legal errors retroactive to the time  they occurred. In
Tennessee Valley Municipal Gas Associa- tion v. FPC, 470 F.2d 446
(D.C. Cir. 1972), we held: "If the  policy of the Natural Gas Act is
not arbitrarily to be defeated  by uncorrected Commission error, the
[injured party] must  be put in the same position that it would have
occupied had  the error not been made." Id. at 452. In Public Service
Co.  of Colorado v. FERC, 91 F.3d 1478 (D.C. Cir. 1996), we  stated:
"Absent detrimental and reasonable reliance, any- thing short of full
retroactivity ... allows [some parties] to  keep some unlawful
overcharges without any justification at  all. The court strongly
resists the Commission's implication  that the Congress intended to
grant the agency the discretion  to allow so capricious a thing." Id.
at 1490. The Public  Service Co. decision was made in the context of
the Natural  Gas Policy Act. We held that the parties were on notice
of a  potential change in the way a tax would be charged to 
customers, and thus did not detrimentally rely on the agen- cy's prior
position. As a result, we held that it was fair to 


make refunds of those tax charges retroactive to the date of  notice.


Finally, Exxon argues that FERC's so-called equitable  exercise of its
discretion failed to give any weight to the  injury to the parties and
the resulting windfall to the Nine  Parties, who benefit because of
agency error, rendering the  agency's ultimate decision irrational.


B. FERC's Position


FERC argues that the Commission properly concluded that  the equitable
approach would be to implement the settlement  on a prospective basis,
as all other TAPS settlements had  been. The cases cited by Exxon
address the issue of whether  FERC is barred from applying a remedy
retroactively, not  whether it is required to do so. FERC's discretion
is at its  zenith when deciding what kind of remedy to apply. See 
Towns of Concord, Norwood, & Wellesley, Mass. v. FERC,  955 F.2d 67,
76 (D.C. Cir. 1992). FERC asserts that it made  its decision based on
several equitable factors6:


FERC took note (1) that parties supported the Nine  Party Settlement
only if it were implemented prospec- tively; (2) that all prior TAPS
cases resolved by settle- ments have been on a prospective basis; (3)
that the  changes adopted by the Settlement Order only modify  limited
aspects of the distillation methodology put in  place in 1993; and (4)
that the TAPS Quality Bank is sui  generis. 81 FERC at 62,467.


FERC Brief at 59. Therefore, FERC argues, it did not  abuse its
discretion. FERC also notes that it did not "bait  and switch" Exxon
in denying the stay because each remedy  must be decided on its own


C. Intervenors' Position


Intervenors note that we have made clear that FERC has  discretionary
authority over whether a settlement should 




__________

n 6 Factor number one, we note, is mentioned only in the agency's 
brief to this court and not in its decision.


have retroactive effect. See CPUC, 988 F.2d at 168. See also  Cities of
Batavia, Naperville, Rock Falls, Winnetka, Geneva,  Rochelle and St.
Charles, Ill. v. FERC, 672 F.2d 64, 85 (D.C.  Cir. 1982) ("It is clear
... that in denying a refund in this case  the Commission also
considered the practical consequences  and the purpose of the Act;
hence we are required to uphold  its exercise of discretionary
power."); Second Taxing Dist.  of the City of Norwalk v. FERC, 683
F.2d 477, 490 (D.C. Cir.  1982) ("Refunds are not mandatory; the
Commission has  discretion to decide whether a refund is warranted in
light of  the interests of the customer and the utility."). OXY did
not  require any result in this case, and in the absence of a clear 
mandate, they argue, FERC properly exercised its discretion.


D. Analysis


We agree that FERC does have a measure of discretion in  determining
when and if a rate should apply retroactively.  However, such
discretion is not without its limits, and we hold  that FERC abused
that discretion.


The agency's passing mention of the filed rate doctrine has  no bearing
on FERC's discretion to reallocate Quality Bank  credits to correct
FERC's erroneous valuations of the distil- late and resid cuts because
all of the TAPS shippers were on  notice as of 1993 that the
valuations were contested. FERC  mentioned the filed rate doctrine not
as a justification for its  exercise of discretion, but in discussing
the prior decision, in  which the filed rate doctrine was decisive. As
we stated in  OXY, "[t]he rule against retroactive ratemaking ...
'does not  extend to cases in which [customers] are on adequate notice
 that resolution of some specific issue may cause a later  adjustment
to the rate being collected at the time of service.'  The goals of
equity and predictability are not undermined  when the Commission
warns all parties involved that a  change in rates is only tentative
and might be disallowed."  64 F.3d at 699 (quoting Natural Gas
Clearinghouse, 965 F.2d  at 1075). In fact, all of the parties
participated in the  proceedings before the agency. Any reliance that
they may  have placed on the rates in light of these proceedings was 
unwarranted. As we stated in Public Service Co., "[a]bsent 


detrimental and reasonable reliance, anything short of full 
retroactivity ... allows [some parties] to keep some unlawful 
overcharges without any justification at all." 91 F.3d at 1490.


There is also a strong equitable presumption in favor of  retroactivity
that would make the parties whole. As we have  stated, "when the
Commission commits legal error, the prop- er remedy is one that puts
the parties in the position they  would have been in had the error not
been made." CPUC,  988 F.2d at 168. This is not to say that FERC must
do so in  every case if the other considerations properly within its 
ambit counsel otherwise. However, FERC's listed equitable  factors
have no bearing on the decision and do not explain its  decision not
to make whole parties who are clearly injured by  undervaluation.
Given the strong presumption in favor of  making injured parties whole
and the incentive that this  creates for the parties to litigate
regarding past errors and  for the agency to correct those errors, on
the record before us  we hold that FERC abused its discretion when it
failed  without adequate explanation to make the revaluation and 
concomitant Quality Bank adjustments retroactive to 1993,  when the
distillation method was adopted.


We recognize FERC's concern that the Nine Parties have  stated that
they would not support the settlement if it applied  retroactively.
However, we cannot uphold on this basis a  contested settlement in
which the settling parties agree to  divvy up a windfall at the
expense of the contesting parties.  The agency cannot simply take a
head count among the  parties in a contested settlement and decide
that since those  who will benefit from a settlement outnumber those
who will  suffer, it is fair to allow the majority to settle the issue
in  their favor. In settlements where the power of the agency is  not
being invoked to overcome the objections of some parties,  all sides
typically give up something to arrive at a mutually  painful but
acceptable position. It should be unsurprising  that the Nine Parties
are unwilling to support the settlement  unless it remains in their
favor if they can invoke the might of  FERC to cram such a settlement
down the minority's throats.  Parties raising legitimate legal
objections cannot be over- looked simply because they are outnumbered,


result is that it sends all parties back to the negotiating table  or
the hearing room. The issue of the effective date of the  new
valuation method is remanded for action consistent with  this


VIII. CONCLUSION


We uphold FERC's decision with two exceptions--we find  that the
decision to use FO-380 less 4.5 cents on the West  Coast and Gulf
Coast Waterborne 3% sulfur No. 6 fuel oil less  4.5 cents on the East
Coast as proxies for the market  valuation for resid was not supported
by substantial evidence  and that the decision to apply the settlement
prospectively  was an abuse of discretion. We vacate those portions of
 FERC's order and remand to the agency to reconsider these  issues in
light of our opinion. We deny the petitions for  review in all other
respects.