Opinions of the United
2006 Decisions                                                                                                             States Court of Appeals
                                                                                                                              for the Third Circuit


9-1-2006

Cloverland Green v. PA Milk Marketing Bd
Precedential or Non-Precedential: Precedential

Docket No. 05-2336




Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2006

Recommended Citation
"Cloverland Green v. PA Milk Marketing Bd" (2006). 2006 Decisions. Paper 402.
http://digitalcommons.law.villanova.edu/thirdcircuit_2006/402


This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova
University School of Law Digital Repository. It has been accepted for inclusion in 2006 Decisions by an authorized administrator of Villanova
University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu.
                                        PRECEDENTIAL

        UNITED STATES COURT OF APPEALS
             FOR THE THIRD CIRCUIT


                      No. 05-2336


    CLOVERLAND-GREEN SPRING DAIRIES, INC.,

                                     Appellant

 THOMAS E. MCGLINCHEY; GERTRUDE GIORGINI;
SUE A. SPIGLER, (“Milk Consumers”) (Intervenors in D.C.)

                           v.

   PENNSYLVANIA MILK MARKETING BOARD;
 BOYD WOLFF*, in his capacity as member of the Board;
      LUKE F. BRUBAKER, Individually and
            as member of the Board;
     BARBARA GRUMBINE*, in her capacity as
              member of the Board;
          BOYD WOLFF, Individually;
       BARBARA GRUMBINE, Individually

 PENNSYLVANIA ASSOCIATION OF MILK DEALERS
             (Intervenors in D.C.)

            (*Amended - See Clerk’s Order dated 6/29/05)
        Appeal from the United States District Court
          for the Middle District of Pennsylvania
           (D.C. Civil Action No. 99-cv-00487)
          District Judge: Honorable Yvette Kane


                    Argued June 8, 2006

               Before: AMBRO, FUENTES
              and NYGAARD, Circuit Judges

             (Opinion filed: September 1, 2006)

Scott T. Wyland, Esquire (Argued)
Kevin J. McKeon, Esquire
Hawke, McKeon, Sniscak & Kennard
100 North Tenth Street
P.O. Box 1778
Garrisburg, PA 17105

      Counsel for Appellant

Wendy M. Yoviene, Esquire
Charles M. English, Jr., Esquire (Argued)
Glenn C. Kennett, Esquire
Thelen, Reid & Priest LLP
701 Eighth Street, N.W.
Washington, D.C. 20001

      Counsel for Appellees

                              

. Warshaw, Esquire
Buchanan Ingersoll & Rooney PC
17 North Second Street, 15th Floor
Harrisburg, PA 17101

        Counsel for Intervenors
        PA Assoc. of Milk Dealers


                 OPINION OF THE COURT


AMBRO, Circuit Judge

        This case, on appeal for the second time to our Court,
concerns whether Pennsylvania’s minimum wholesale prices for
fluid milk violate the dormant Commerce Clause of our
Constitution. In our first opinion, we held that the District Court
improperly granted summary judgment to the Pennsylvania Milk
Marketing Board (“Board”),1 and remanded to the District Court
for a trial. See Cloverland-Green Spring Dairies, Inc. v. Pa.
Milk Mktg. Bd., 

(3d Cir. 2002) (hereafter
“Cloverland I”). The District Court then conducted a six-day


    1
       Although the members of the Board have been sued
individually in their official capacities, we refer to them as the
“Board”. We also note that the Pennsylvania Association of
Milk Dealers intervened as a defendant in the District Court in
support of the Board, and filed an appellee’s brief on this appeal.
For simplicity, we shall refer to all appellees as the “Board”.

                                3
bench trial and ruled in favor of the Board. We now affirm.

                       I. Milk Economics

       Before addressing the facts of this case, we believe it
helpful (if not necessary) to explain how milk pricing works.
On the surface, the path of milk from the cow to the consumer
seems simple. A farmer (commonly called a “producer”) sells
raw milk to a processor (or “handler”). The handler processes
the raw milk into fluid milk or other dairy products, packages it,
and sells the product to a retailer. The retailer, in turn, sells to
the consumer.

        Yet this apparent simplicity is deceptive. Federal and
state regulations regarding the sale of milk make “Byzantine” an
apt, and none too pejorative, description. See, e.g., Lansing
Dairy, Inc. v. Espy, 

, 1344 (6th Cir. 1994) (“[T]he
system established by [federal law] to regulate the sale of milk
is of labyrinthine complexity.”); Kenneth W. Bailey, Marketing
and Pricing of Milk and Dairy Products in the United States 109
(1997) (noting that the rules governing milk prices are “mind-
boggling”);2 Jim Chen, Around the World in Eighty Centiliters,


  2
     We note that Professor Bailey testified as an expert witness
for the Board in the District Court. We do not, of course, cite
his textbook herein as a substitute for his testimony, but rather
as a reference for the workings of the federal milk pricing
system, which the parties do not dispute.

                                 4
15 Minn. J. Int’l L. 1, 6 (2006) (describing the federal milk
pricing structure as “outlandishly complex”).3 This complexity,

   3
      As Judge Jerome Frank of the United States Court of
Appeals for the Second Circuit once observed (in terms more
gilded than our era musters):

             [T]he ‘milk problem’ is exquisitely
             complicated. The city-dweller or
             poet who regards the cow as a
             symbol of bucolic serenity is
             indeed naive. From the udders of
             that placid animal flows a bland
             liquid indispensable to human
             health but often provoking as much
             human strife and nastiness as strong
             alcoholic beverages. . . . The[se]
             difficulties have given rise to much
             legislation and are reflected in
             many judicial decisions. . . . The
             milk problem is so vast that fully to
             comprehend it would require an
             almost universal knowledge
             ranging from geology, biology,
             chemistry and medicine to the
             niceties of the legislative, judicial
             and administrative processes of
             government. It affects an industry
             immense in scope, for dairying is
             [one of] the largest . . . branch[es]
             of agriculture in this country.

                              5
and the large sums of money at stake,4 have spawned an
extraordinary amount of litigation. It has been estimated that
“no other commodity in the United States has been involved in
as many legal challenges in regard to how it is marketed” as
milk. 

.

       Two systems of milk price regulation concern us on this
appeal: the federal system and Pennsylvania’s state system. We
describe each in turn.

       A.     Federal Milk Price Regulation

              1.     Unique Nature of Milk

       Milk is a unique agricultural commodity. The Supreme
Court has recognized “two distinctive and essential phenomena
of the milk industry[:] a basic two-price structure that permits a
higher return for the same product, depending on its ultimate
use, and the cyclical characteristic of production.” Zuber v.
Allen, 

, 172 (1969); see also Lehigh Valley
Farmers v. Block, 

, 411 (3d Cir. 1987) (same). The


Queensboro Farms Prods. v. Wickard, 

, 974-75 (2d
Cir. 1943).
   4
      In 2003, for example, U.S. dairy farmers sold about 19.7
billion gallons of milk for more than $21 billion. Gen.
Accounting Office, Dairy Industry: Information on Milk Prices,
Factors Affecting Prices, and Dairy Policy Options 1 (2004).

                                6
first phenomenon derives from the fact that raw milk may be put
to two general uses by handlers: fluid milk for drinking and
manufactured dairy products such as cheese, butter, and ice
cream. See 

. Fluid milk is more
expensive for handlers to produce and market because it is
highly perishable (a gallon of milk must ordinarily be marketed
to the consumer within two days of milking), requires more
sanitation and processing than other dairy products, and is
heavier and thus more expensive to transport. See 

n.3;

4, 46, 109-10; Alden C. Manchester & Don P.
Blayney, Econ. Research Serv., U.S. Dep’t of Agric., Milk
Pricing in the United States 2 (2001), available at
http://www.ers.usda.gov/publications/aib761/aib761.pdf. Fluid
milk for drinking therefore costs the consumer more per pound
than milk that has been manufactured into other dairy products,
and the partial inelasticity of demand for fluid milk (viewed as
a dietary staple by many Americans) means profit margins on
fluid milk sales may also be higher. See 

4;
see also Smyser v. Block, 

, 516 (3d Cir. 1985)
(“Milk that is ultimately used for fluid purposes has traditionally
commanded a higher price than milk of the same grade and
quality used for manufactured products. This difference is not
entirely accounted for by differences in cost.”).

      Farmers also produce two “grades” of marketable raw
milk. Grade A milk is fit for drinking, and therefore can be
processed into fluid milk or manufactured products, while Grade
B milk is not fit for drinking but may be used in other dairy


                                7
products. Manchester & 

; David L. Baumer,
Federal Regulation of Milk Production and Sale is Growing at
the Expense of State Authority, 12 J. Agric. Tax. & L. 36, 38-39
(1990). Not surprisingly, in an entirely unregulated market,
handlers would have an incentive to produce manufactured dairy
products using Grade B milk and reserve as much Grade A milk
as possible for the more lucrative fluid milk market. At the
same time, farmers would want to be paid a higher price for
their Grade A milk than for Grade B milk, but handlers would
want to pay less for Grade A milk they intended to put to use in
manufactured dairy products (i.e., surplus Grade A milk they
cannot sell as fluid milk).

       The complexity deepens when we consider the second
phenomenon identified in Zuber: the cyclical nature of milk
production. Demand for fluid milk, though somewhat inelastic
with respect to price, is temporally cyclic, with demand higher
in the fall and winter months than in the spring and summer
months. See 

-73; 

4; see
also John R. Snyder, A Summary: Political and Economic
Analysis of Milk Marketing, 1980-81 Agric. L.J. 297, 310-11.
Unfortunately, cows do not work this way, and produce more
milk in the spring and summer (known as the “flush” season)
than in the fall and winter (the “short-supply” season).
Manchester & 

; 

.
Thus, “it [is] necessary to coordinate a supply that is rising when
fluid milk demand is falling.” Manchester & 

. Otherwise, handlers would attempt to “take advantage of this


                                8
surplus to obtain bargains [from producers] during glut periods,”
which would lead farmers to increase production to maintain a
steady income, “and the disequilibrium snowballs.” 

; 

(“In an unregulated market
‘cutthroat’ competition for more profitable fluid milk sales can
lead to an overall decline in prices.”); see also 

(explaining that the unique characteristics of the milk
industry, if unregulated, lead to “chaotic marketing conditions”).

              2.      The Federal System

       Federal regulation of the nation’s dairy industry began in
earnest in the 1930s, when falling prices caused by the Great
Depression led to “utter chaos” in the milk market. 

. The first attempt at regulation was the Agricultural
Adjustment Act (“AAA”) of 1933, 48 Stat. 31, which
empowered the Secretary of Agriculture to promulgate
emergency licensing requirements in the dairy industry to
regulate output and price. See 

-75. The
Supreme Court’s decision in A.L.A. Schechter Poultry Corp. v.
United States, 

(1935), however, struck down a
similarly broad delegation of rulemaking authority under the
National Industrial Recovery Act of 1933, 48 Stat. 195, and
Congress acted to protect regulation of the dairy market by first
amending the AAA and then enacting the Agricultural
Marketing Agreement Act (“AMAA”) of 1937, 50 Stat. 246,
which replaced the emergency measures with a permanent
system of milk regulation. See 

;


                                9
Defiance Milk Prods. Co. v. Lyng, 

, 1066-67 (6th
Cir. 1988).

       Under the AMAA, the Secretary of Agriculture is
empowered to regulate the nation’s milk markets by issuing
“orders” that regulate prices in defined geographic areas. See 7
U.S.C. § 608c(1); 

; Lehigh 

. These orders are meant to ensure “a sufficient
quantity of pure and wholesome milk to meet current needs and
further to assure a level of farm income adequate to maintain
productive capacity sufficient to meet anticipated future needs.”
7 U.S.C. § 608c(18). Although their number has varied over the
years, there are currently ten federal milk marketing order
areas.5 See 7 C.F.R. pts. 1001-1135; U.S. Dep’t of Agric.,


  5
     These are the Northeast, Appalachian, Southeast, Florida,
Mideast, Upper Midwest, Central, Southwest, Arizona, and
Pacific Northwest orders.        See U.S. Dep’t of Agric.,
Consolidated Milk Marketing Order Areas, available at
http://www.ams.usda.gov/dairy/dymap.htm. There are also
several parts of the country that are not regulated by a federal
order, including parts of Idaho, New York, Pennsylvania, and
North and South Dakota, and most or all of California, Maine,
Montana, Nevada, Utah, Virginia, and Wyoming. 

orders are often referred to by a number
that corresponds to the implementing regulation that defines the
order. So, for example, the Northeast order (encompassing all
or part of Vermont, New Hampshire, Massachusetts, Rhode
Island, Connecticut, New York, Pennsylvania, New Jersey,

                               10
Consolidated Milk Marketing Order Areas, available at
http://www.ams.usda.gov/dairy/dymap.htm.          Within these
orders, the Secretary sets prices at which raw milk may be sold.
There are two main features of this pricing structure that merit
explanation.

          Handlers pay for their milk according to a classified
pricing arrangement. See 7 U.S.C. § 608c(5)(A). Grade A raw
milk is classified according to its end use. Class I milk is fluid
milk for drinking, while other classes (II, III, and IV) are
assigned to raw milk used in various manufactured products. 7
C.F.R. § 1000.40; see West Lynn Creamery, Inc. v. Healy, 

, 189 n.1 (1994). It is not necessary for our purposes to
explain in detail the pricing mechanism for each class, see, e.g.,
7 C.F.R. § 1000.50; we simply note that handlers pay more for
Grade A raw milk they intend to put to use as Class I milk than
they do for Grade A raw milk they intend to put to use as Class
II, III, or IV milk.6 See Lansing 

; Smyser,


Delaware, Maryland, and the District of Columbia, as well as a
few counties in Northern Virginia) is defined in 7 C.F.R. part
1001, and is known commonly as “Order 1”. The Mideast order
(encompassing all or part of Michigan, Indiana, Ohio, Kentucky,
West Virginia, and Pennsylvania) is defined in 7 C.F.R. part
1033, and is known commonly as “Order 33”.
    6
      To be clear, however, the different classes are merely
administrative distinctions to take account of the fact that
handlers charge different wholesale prices for fluid milk and

                               

.

        Producers do not, however, receive more or less money
for raw milk based on the handler’s use. Instead, producers
receive a uniform “blend” price that is essentially a weighted
average of the price of all classes of milk sold in the region. See
7 U.S.C. 608c(5)(B); 

; Lehigh 

. This system “eliminate[s] the potential
situation where one producer, who sold milk to a fluid
processor, would get a higher milk price than his neighbor who
sold to a cheese plant.” 

; see also Lehigh

-12 (“By mandating that producers
within a market area receive a uniform price based on the use of
milk within the area, the [AMAA] eliminates the incentive for
dairy farmers to attempt to compete with their neighbors through
lowering their prices.”).

        Of course, this would ordinarily lead to a situation in
which handlers dealing mostly in lower classes of milk
essentially subsidize those dealing mostly in Class I milk,
because the blend price paid to the producer will be more than,
say, the classified price of Class IV milk, but lower than the
classified price of Class I milk. To compensate for this, the
regulations establish a “Producer Settlement Fund,” into which



manufactured products. The raw milk is all Grade A, and is
alike in all respects except the artificially created class
distinction that depends on the use to which the milk is put.

                                12
handlers who have underpaid contribute an amount equal to the
difference between the blend price and classified price, which is
then distributed to handlers who have overpaid. See 7 C.F.R.
§ 1000.70; Lansing 

; 

.

       By charging handlers less for Grade A raw milk used in
storable dairy products like cheese and butter than for raw milk
used as Class I drinking milk, while assuring that producers are
paid the same no matter the use to which the raw milk is put,
handlers have an incentive to use surplus Grade A milk in
manufactured products during the glut months, and producers
maintain a more stable income year-round. This, in turn, guards
against overproduction in the flush season. Moreover, the
Producer Settlement Fund relieves pressures on handlers to deal
exclusively in fluid milk to the detriment of other dairy
products.

       B.     Pennsylvania Milk Price Regulation

       Although state control of milk markets has declined
significantly in the wake of federal regulation, particularly
through court challenges, it survives in a few states, including
Pennsylvania. See 

, 210; see generally

. Parts of Pennsylvania are federally regulated —
the southeast and southcentral parts of the Commonwealth are
under the Northeast order, and the western part of the
Commonwealth is under the Mideast order — while other parts


                               13
are not regulated by the federal system. Pennsylvania’s system
covers the entire Commonwealth, however.

       Under Pennsylvania’s Milk Marketing Law, 31 Pa. Cons.
Stat. § 700j, the Board is granted “power to supervise,
investigate and regulate the entire milk industry of th[e]
Commonwealth, including the production, transportation,
disposal, manufacture, processing, storage, distribution,
delivery, handling, bailment, brokerage, consignment, purchase
and sale of milk and milk products in th[e] Commonwealth.” 31
Pa. Cons. Stat. § 700j-301. The statute requires that all handlers
operating in Pennsylvania be licensed by the Board, see 

and empowers the Board to set minimum prices paid
to producers, handlers, and retailers to “benefi[t] . . . the public
interest [and] best protect the milk industry of the
Commonwealth and insure a sufficient quantity of pure and
wholesome milk to [its] inhabitants.” 

       In-state handlers who buy raw milk from in-state
producers for in-state use must pay “over-order” prices — which
include a premium above the applicable Class I milk prices —
to assure producers a “reasonable return” based on the
“conditions affecting the milk industry in each marketing area.”7


  7
    According to the Board, “the state-mandated premium has
resulted in over 335 million additional dollars being paid to
Pennsylvania farmers” since 1988. Pa. Milk Mktg. Bd., Fiscal
Year Report, 2003-2004 10 (2004), available at http://

                                14
Id.; see also 

(“The [B]oard shall fix, by official
order, the minimum prices or a formula for setting of minimum
prices to be paid by milk dealears or handlers to producers for
milk or milk components sold or delivered or made available on
consignment or otherwise by producers to dealers or handlers.”).
The over-order premium only applies to sales of Class I milk.
See Pa. Milk Mktg. Bd., Order No. A-913 (2001) (establishing
the method for calculating the over-order premium, and
specifying that the premium “shall be based on milk that is
produced, processed, and utilized as Class I milk in
Pennsylvania”).8

       Pennsylvania also exports raw milk to other states. The
over-order price, however, does not apply to sales to out-of-state
handlers.



www.mmb.state.pa.us/mmb/lib/mmb/2003-2004.pdf. From July
2003 to June 2004, the over-order premium resulted in
Pennsylvania farmers receiving $30.8 million more than they
would otherwise have received. 

    8
        The Board has promulgated a complex mathematical
formula for calculating the over-order premium. See Pa. Milk
Mktg. Bd., Order No. A-913 (2001). It is not necessary to
explain that process here. As of this writing, in all Pennsylvania
milk marketing areas handlers must add $1.60 per
hundredweight (approximately 11.6 gallons) to the price of
Class I milk. See Pa. Milk Mktg. Bd., Order No. A-938 (2006).


                               15
        All handlers who sell milk to retailers in Pennsylvania
must do so at minimum wholesale prices that also guarantee a
“reasonable return,” which state law defines as between 2.5%
and 3.5% of net sales. 31 Pa. Cons. Stat. § 700j-801; see also 

(“The [B]oard shall fix, by official order . . ., the
minimum wholesale and retail prices, and may fix, by official
order, the maximum wholesale and retail prices, to be charged
and received by milk dealers or handlers for milk sold,
delivered, handled or consigned within any milk marketing area
of the Commonwealth.”).

        The Board has divided Pennsylvania into six milk
marketing areas. Areas 1 and 4 (the Areas at issue in this
appeal) consist of the counties in southeast and southcentral
Pennsylvania that are also covered by the federal Northeast
order. Area 5 consists of the counties in western Pennsylvania
that are also covered by the federal Mideast order. The
remaining Areas (2, 3, and 6) consist of the counties that are not
regulated by the federal order system. See Pa. Milk Mktg. Bd.,
Fiscal Year Report, 2003-2004 10 (2004), available at
http://www.mmb.state.pa.us/mmb/lib/mmb/2003-2004.pdf. The
Board sets minimum wholesale and retail prices, as well as over-
order premiums, in each Area based on the circumstances of that
region, and enforces the minimum prices by regular audits of
handlers’ books. 

sells milk (raw or processed)
at a price below the mandatory minimums set by the Board is
subject to criminal penalties, including fines or imprisonment.
31 Pa. Cons. Stat. § 700j-1001 & 1002.


                               16
       The United States Department of Agriculture (“USDA”)
has identified several important differences between the
Pennsylvania system and the federal orders. First, Pennsylvania
divides milk into only two classes when calculating the price
handlers must pay, as opposed to the four classes applicable to
prices under federal orders. See Milk in the New England and
Other Marketing Areas, 64 Fed. Reg. 16,026, 16,056 (Apr. 2,
1999). As in the federal system, Class I milk in Pennsylvania is
primarily fluid milk for drinking, but Class II milk encompasses
all other dairy products. Second, the location where milk is
distributed for sale determines the applicable minimum price in
Pennsylvania, while under federal orders the price is determined
based on the handler’s plant location. 

prices
paid to farmers under the Pennsylvania system are calculated
based on the individual handler’s usage, and not the use of the
various milk classes in the wider market. Id.; see also 7 Pa.
Code § 143.11 (“Dealers shall pay their producers on a weight
and butterfat basis as determined by the utilization of the milk
received at each plant or receiving station.”). This includes, of
course, the over-order premium, which is added to the Class I
price and then (as part of the Class I price) blended with the
Class II price to yield the handler’s “plant blend” price paid to
its producers.

              II. Facts and Procedural History

       With this background in mind, we proceed to the facts of
this case. Appellant Cloverland-Green Spring Dairies, Inc.


                               17
(“Cloverland”) is a milk handler located in Maryland. It buys
approximately 90% of its raw milk from Pennsylvania
producers, but does not sell processed fluid milk in
Pennsylvania. In March 1999, Cloverland brought suit against
the Board’s members in their official capacities under 42 U.S.C.
§ 1983, alleging that Pennsylvania’s milk pricing scheme
eliminates competition based on price and, as such, nullifies
Cloverland’s competitive advantages, thereby effectively
barring it from Pennsylvania’s market. This, it argues, violates
the Constitution’s dormant Commerce Clause (described in Part
III.A below). Three individual milk consumers in Pennsylvania
intervened as plaintiffs in the suit, and the Pennsylvania
Association of Milk Dealers intervened as a defendant in
support of the minimum prices.

       A.     Summary Judgment and Appeal

         The District Court granted summary judgment to the
Board, see Cloverland-Green Spring Dairies, Inc. v. Pa. Milk
Mktg. Bd., 

(M.D. Pa. 2001), and
Cloverland appealed. We affirmed the grant of summary
judgment with respect to Pennsylvania’s minimum retail prices
(i.e., prices charged to consumers, which are not at issue on this
appeal), but reversed with respect to the minimum wholesale
prices. Cloverland 

. Viewing all facts in the
light most favorable to Cloverland — as required at that stage of
the proceedings, see 

held that a reasonable trier of fact
could find (on the facts presented) that Pennsylvania’s minimum


                               18
wholesale milk prices were unconstitutional, thus precluding
summary judgment for the Board. Specifically, we noted that
Cloverland indicated it had evidence that it could offer prices
below Pennsylvania’s wholesale floor. Although, based on the
limited record available at that stage of the proceedings, it was
not entirely clear to us why Cloverland could offer lower prices,

we observed that Cloverland (as an out-of-state
handler) is exempt from paying the over-order premium on
Class I milk purchased from Pennsylvania producers, and thus
its raw milk costs could be lower than similarly situated
Pennsylvania handlers. 

As such, “a reasonable trier
of fact could find that Pennsylvania’s minimum wholesale prices
eliminate a competitive advantage enjoyed by out-of-state
dealers like Cloverland,” which would trigger heightened
scrutiny under the dormant Commerce Clause. 

Moreover, there was no evidence (at that time) that there were
Pennsylvania firms more efficient than Cloverland, and thus the
minimum price floors did not necessarily affect all in-state and
out-of-state firms equally. Id.9


  9
      Indeed, on the facts Cloverland offered, it appeared to us
that “[b]y calibrating wholesale price floors for a particular milk
marketing area to the operating costs of average dealers in that
area, the Commonwealth enables Pennsylvania dealers to
operate less efficiently without fearing losses to lower-cost
competitors like Cloverland. This aspect of Pennsylvania’s
scheme appears to run afoul of the cardinal rule that states may
not shield in-state businesses from out-of-state competitors.”
Cloverland 

n.17.

                                19
         Even if the District Court were to conclude that
heightened scrutiny did not apply, we held that “the record . . .
amply supports a finding that the wholesale floors
. . . incidentally burden interstate commerce by making it more
difficult for out-of-state dealers to attract new business in a
market dominated by in-state dealers,” 

and thus the
District Court would have to employ the balancing test set forth
in Pike v. Bruce Church, Inc., 

(1970), to determine
whether the burdens on interstate commerce substantially
outweighed the putative local benefits. Cloverland 

. We held that the “the purpose asserted in the
[Pennsylvania Milk Marketing Law’s] text deserve[d] little
deference” in identifying the law’s benefits because no
Pennsylvania milk processors objected to the minimum prices
(indeed, the Pennsylvania Association of Milk Dealers
intervened in support of the law), and the price scheme therefore
“appear[ed] to [burden] only . . . out-of-state dealers.” 

Moreover, we found only “meager evidence” that the
minimum prices were necessary to sustain Pennsylvania’s dairy
industry — since “the federal producer price floors [appear to]
provide ample protection against predatory pricing” in other
states, and the Pennsylvania dairy industry was quite successful
(indeed, it exported most of its raw milk) — and thus it seemed
to us unlikely that the minimum prices were necessary to secure
a supply sufficient to meet the needs of Pennsylvania’s
residents. 

Moreover, the burdens on interstate
commerce appeared substantial: Cloverland’s stated evidence
suggested it was “virtually impossible to displace incumbent


                               20
dealers in Pennsylvania without offering prices below the
Board-mandated floors.” 

Thus, construing all the
facts in the light most favorable to Cloverland, the
Commonwealth’s minimum wholesale milk prices could fail the
Pike balancing test.

       C.     Second District Court Proceeding

       On remand, the District Court permitted full discovery
and held a bench trial at which numerous witnesses testified.
Based on this evidence, the Court made detailed findings of fact
that largely contradicted the facts (viewed in the light most
favorable to Cloverland) on which we relied in our earlier
decision. The District Court found that Cloverland — which,
again, buys about 90% of its raw milk from Pennsylvania
suppliers — purchases about 35% of this milk from dairy farm
cooperatives, which negotiate prices with Cloverland equal to
the over-order price charged to Pennsylvania processors.
Cloverland-Green Spring Dairies, Inc. v. Pa. Milk Mktg. Bd.,
No. 1:CV-99-487, slip op. at 5 (M.D. Pa. 2005) (hereafter
“Cloverland II”). The other 65% is purchased from independent
Pennsylvania farmers, who do not charge the over-order
premium to Cloverland because, as an out-of-state handler, it is
exempt. 

The Board nonetheless offered expert
testimony that Cloverland actually had higher raw milk costs
and was less efficient than three of four sample dairies, which
the Court accepted because Cloverland did not produce



                              21
sufficient countervailing evidence.10 

15. Thus, the
Court found that “[a]lthough Cloverland can sell milk at a cost
below the minimum wholesale price floor established by the
Board, Cloverland does not have an out-of-state price or
efficiency comparative advantage.” 

       Indeed, the Court concluded that, even if Cloverland’s
exemption from the over-order premium conferred an advantage
because it is an out-of-state firm, that advantage “arises from the
very Milk Law pricing structure [Cloverland] seeks to partially
invalidate,” and thus, if Cloverland’s suit were successful, its
advantage would be “ephemeral.” 

In the Court’s
view, the invalidation of the wholesale price would necessarily
lead to the abandonment of the over-order premium, which
would make it economically unfeasible for small dairy farms to
remain in business and thus would force them into cooperatives
(which negotiate prices with out-of-state processors equal to the
over-order price). 

       The District Court also found that, although “[m]ost
processors sell at or around the wholesale floor price, [thus]
precluding appreciable amounts of competition based upon
price,” 

Cloverland failed to show that competition on
non-price factors did not exist. Indeed, the Court found


  10
       The District Court expressly found the Board’s expert on
milk pricing to be more credible than Cloverland’s. Cloverland
II, slip op. at 6 n.4.

                                22
“competition among processors in quality, service, and the
ability to bundle the delivery of non-milk based drinks,” and
“[a]lthough price is a significant factor in competing for
business within the Pennsylvania dairy market, twice as many
retailers choose name recognition as the number one factor
informing their choice of supplier.” 

Thus, even if
Cloverland could compete on the basis of price, the Court found
that it failed to prove it could penetrate the Pennsylvania market.
Id.11 Indeed, the Court noted that Cloverland only attempted


   11
       At one point in its opinion, the District Court stated that
“Cloverland has established that it can compete with
Pennsylvania dealers only by selling milk at prices below the
wholesale milk floor,” Cloverland II, slip op. at 7 (emphasis
added), but we do not believe the Court meant what it literally
said. In the same paragraph, the Court explained that robust
competition exists in the Pennsylvania market on a variety of
non-price factors, and “even were Cloverland able to establish
that it could sell milk for the lowest price, [it] would not enjoy
a competitive advantage that would enable it to penetrate the
Pennsylvania dairy market.” 

The Court
expressly found “that the minimum price floor does not have the
practical effect of preventing Cloverland from entering the
Pennsylvania milk market,” 

and later
observed that Cloverland “offered no credible evidence that it
could in fact compete in the Pennsylvania marketplace, absent
of the Milk law.” 

        Under these circumstances, we are confident that the
District Court, by typographical error or otherwise, misstated the
intended proposition when it said Cloverland did establish the

                                23
(and failed) to solicit business from three retailers in Area 4 and
none in Area 1, and provided no evidence that these retailers
rejected it because of price. 

20.

        Thus, the District Court concluded that Cloverland failed
to prove that the Pennsylvania Milk Marketing Law nullified
any competitive advantage it enjoyed by virtue of being an out-
of-state firm, and thus heightened scrutiny did not apply.

       The Court then considered whether the minimum pricing
scheme’s incidental burdens on interstate commerce
substantially outweighed the putative local benefits, as required
by the Pike balancing test. It concluded that minimum
wholesale prices confer substantial benefits on Pennsylvania’s
dairy industry, because the federal minimums “ha[ve] not
adequately guaranteed enough income to protect Pennsylvania
dairy farms since 1988,” when federal deregulation lowered the
minimums. 

26. Minimum wholesale prices allow
handlers a sufficient return to enable them to pay the over-order
premium, which in turn allows Pennsylvania’s independent
dairy farms to survive. 

24. Based largely on evidence
produced at trial regarding the abolition of minimum prices in
California, the District Court concluded that if Pennsylvania’s
market were dominated by a few large players, “[c]onsumer
prices would fall in the short term, but rise heavily in the long
term.” 

25. And, if small farms went out of business,


very thing it repeatedly explained Cloverland did not establish.

                                24
the Court concluded that the “agricultural infrastructure” that
supports the dairy industry — including “dealers, feed stores,
veterinarians, and other business[es]” — might suffer.12 

26.

        In contrast, the Court found only incidental burdens on
interstate commerce: the minimum prices simply placed
Cloverland on the same footing as every other handler doing
business in Pennsylvania, did not prevent interstate competition
on non-price factors, and the dominance of incumbent handlers
could be attributed to Pennsylvania retailers’ preference for
doing business with local handlers “independent of regulatory
pressures.” 

Thus, the Court held that, under the Pike
test, Cloverland failed to prove a violation of the dormant
Commerce Clause, and granted judgment in the Board’s favor.
Cloverland appeals.13


  12
     These conclusions were based largely on the testimony of
the Board’s experts, which the District Court expressly found
more credible than Cloverland’s expert. Cloverland II, slip op.
at 27.
  13
     The District Court had subject matter jurisdiction over this
case pursuant to 28 U.S.C. §§ 1331 and 1332, and we have
jurisdiction over the appeal pursuant to 28 U.S.C. § 1291. Our
review of the District Court’s decision is subject to familiar
standards: we review findings of fact for clear error (meaning
we will reverse only if we are “left with a definite and firm
conviction that a mistake has been committed”), and exercise

                               25
                          III. Analysis

         Cloverland raises two issues on appeal. It contends its
competitive advantages require heightened scrutiny, regardless
whether those advantages derive from its status as an out-of-
state firm, and that the mandatory minimum wholesale prices are
the only reason it cannot compete in the Pennsylvania market.
Even if heightened scrutiny does not apply, Cloverland argues
that under the Pike test the burdens on interstate commerce
clearly outweigh any putative benefits.

       A.     The Dormant Commerce Clause

       Under the federal Constitution’s Commerce Clause,
Congress has explicit power to “regulate Commerce . . . among
the several States.” U.S. Const. art. I, § 8, cl. 3. This clause
also has an implied requirement (often called the “negative” or
“dormant” aspect of the clause) that states not “mandate
differential treatment of in-state and out-of-state economic
interests that benefits the former and burdens the latter.”
Granholm v. Heald, 

, 472 (2005) (internal
quotation marks omitted); West Lynn 

(same). Under the dormant Commerce Clause, it is
“[a]xiomatic . . . that a state cannot impede free market forces to
shield in-state businesses from out-of-state competition.”


plenary review over legal conclusions. Gordon v. Lewiston
Hosp., 

, 201 (3d Cir. 2005).

                                26
Cloverland 

. This includes “forcing [out-of-
state businesses] to ‘surrender whatever competitive advantages
they may possess’” as the price of doing business in the state.

Distillers Corp. v. N.Y. State Liquor
Auth., 

, 580 (1986)).

       In considering whether a state law violates the dormant
Commerce Clause, the inquiry is twofold: a court considers first
whether “heightened scrutiny” applies, and, if not, then
considers whether the law is invalid under the Pike balancing
test. See, e.g., Cloverland 

-11.

       Heightened scrutiny applies when a law “discriminates
against interstate commerce” in its purpose or effect. C & A
Carbone, Inc. v. Town of Clarkstown, 

, 390 (1994);
Harvey & Harvey, Inc. v. County of Chester, 

, 797-
98 (3d Cir. 1995).14 The party challenging the statute has the
burden of proving the existence of such discrimination, Harvey
& 

, and the burden then shifts to the state
to prove that “the statute serves a legitimate local purpose, and
that this purpose could not be served as well by available
nondiscriminatory means.” Maine v. Taylor, 

, 138
(1986); Harvey & 

(same). This standard


   14
       Notably, “in order to find a dormant Commerce Clause
violation there is no requirement that discrimination must be the
‘primary’ purpose or effect” of the challenged state law. Harvey
& 

.

                               27
renders all but the most unusual statute invalid. 

; 

(“When a state
statute directly regulates or discriminates against interstate
commerce, or when its effect is to favor in-state economic
interests over out-of-state interests, we have generally struck
down the statute without further inquiry.”); Cloverland 

-11; see also Harvey & 

(referring to this standard as “strict scrutiny”).

        There are two general types of discrimination that a
plaintiff may show to trigger heightened scrutiny (although they
are not entirely distinct, and overlap in many ways). First, it
may show that the challenged state statute has extraterritorial
effects that adversely affect economic production (and hence
interstate commerce) in other states, thereby forcing “producers
or consumers in other States [to] surrender whatever competitive
advantages they may possess” to “give local consumers an
advantage over consumers in other states.” 

; see Baldwin v. G.A.F. Seelig, Inc., 

,
521, 527 (1935) (“New York has no power to project its
legislation into Vermont by regulating the price to be paid in
that state for milk acquired there,” because such regulation
“set[s] up what is equivalent to a rampart of customs duties
designed to neutralize advantages belonging to the place of
origin.”).

      Second, it may show that the “object [of the law] is local
economic protectionism,” in that it disadvantages out-of-state


                              28
businesses to benefit in-state ones. 

;
see West Lynn 

(holding that a
Massachusetts state law requiring milk processors, including
out-of-state firms, to pay a premium to a state fund that was then
disbursed only to in-state producers was “effectively a tax which
makes milk produced out-of-state more expensive,” thus
discriminating against out-of-state milk); Hunt v. Wash. State
Apple Adver. Comm’n, 

, 351-53 (1977)
(overturning a North Carolina statute that forbade importing
apples bearing a quality mark other than one issued by the
USDA, because Washington’s apples bore a state quality mark
and thus “the statute’s consequence [is] raising the costs of
doing business in the North Carolina market for Washington
apple growers and dealers, while leaving those of their North
Carolina counterparts unaffected”); Polar Ice Cream &
Creamery Co. v. Andrews, 

, 375-77 (1964)
(holding that a Florida statute requiring in-state milk processors
to purchase raw milk from in-state producers unlawfully
discriminated against interstate commerce).15


   15
      The Board argues that the Supreme Court’s decision in
Pharmaceutical Research and Manufacturers of America v.
Walsh, 

(2003), stands for the proposition that two
“separate and distinct” types of heightened scrutiny analysis
exist (extraterritorial effect and discrimination), and that
“extraterritorial effect” cases like Baldwin and Brown-Forman
are not relevant to “discrimination” cases like Carbone and
Harvey, and vice versa. Pa. Milk Mktg. Bd. Br. at 49-51.
       The Board substantially overreads Pharmaceutical

                               29
        In deciding whether a state law discriminates against out-
of-state businesses, it is immaterial whether the statute or
ordinance also burdens some in-state businesses. See 

(“The ordinance is no less discriminatory
because in-state or in-town processors are also covered by the
prohibition.”); Cloverland 

(noting that
Carbone “‘explicitly rejected the argument that a disputed
statute would have to favor all in-state businesses as a group —


Research. The Supreme Court discussed the “extraterritorial
effect” cases apart from other discrimination cases only because
the petitioner framed its argument that way. See Pharm.
Research, 

. Nowhere did the Supreme Court
suggest that the two lines of cases are mutually exclusive;
indeed, so-called “discrimination” cases like Carbone, West
Lynn Creamery, and Washington State Apple rely heavily on
so-called “extraterritorial effect” cases like Baldwin, and vice
versa. As we noted in Cloverland I, both types of cases may
share the common element of “state laws that are facially neutral
but have the effect of eliminating a competitive advantage
possessed by out-of-state firms, [thus] trigger[ing] heightened

(citing Wash. State Apple, 

, and Baldwin, 

); see also 

n.14
(“[T]he problem with the law invalidated in Baldwin was not
merely its extraterritorial reach, but that it had the practical
effect of discriminating against out-of-state milk producers by
eliminating their competitive advantage.”). Thus, while it may
sometimes be useful to consider the two types of cases
separately for ease of analysis, they are actually just two forms
of discrimination, with significant overlap.

                               30
a statute may be invalid if it favors only a single or finite set of
businesses’” (quoting Harvey & 

)).

        If the plaintiff does not succeed in showing that the
purpose or effect of the state law discriminates against interstate
commerce — but, rather, the statute “regulates even-handedly
to effectuate a legitimate local public interest, and its effects on
interstate commerce are only incidental” — the court must
determine whether “the burden imposed on such commerce is
clearly excessive in relation to the putative local benefits.” 

; Cloverland 

(same). This
balancing test is significantly less restrictive of state laws than
heightened scrutiny.

       Often, “there is no clear line separating the category of
state regulation that is virtually per se invalid under the
Commerce Clause, and the category subject to the Pike v. Bruce
Church balancing approach.” 

;
see Cloverland 

(“[S]ometimes the distinction
between state laws subject only to Pike balancing and those that
are nearly per se invalid is ‘hazy.’”). In determining whether
heightened scrutiny should be applied instead of the Pike test,
“the critical consideration is the overall effect of the statute on
both local and interstate activity,” 

, with special attention paid to whether a “facially neutral”
state law “ha[s] the effect of eliminating a competitive
advantage possessed by out-of-state firms” (which “trigger[s]
heightened scrutiny”). Cloverland 

.


                                31
        B.     Heightened Scrutiny

         Cloverland’s argument against the District Court’s
heightened scrutiny analysis may be distilled as follows. The
District Court concluded that Cloverland has some advantages
over some Pennsylvania firms — it is more efficient than some,
it sells milk in Maryland for less than the minimum wholesale
prices that exist in Pennsylvania, and its exemption from the
over-order premium for raw milk purchased from independent
Pennsylvania dairy farmers means its raw milk costs might
theoretically be lower than its Pennsylvania competitors’ costs.
Although these advantages may not derive necessarily from
Cloverland’s status as an out-of-state handler,16 the argument
proceeds, there is no question that Pennsylvania’s minimum
prices level the playing field with Pennsylvania firms that have
higher costs and are less efficient, thus nullifying Cloverland’s
alleged competitive advantages. Since it is an out-of-state firm,
Cloverland contends the pricing scheme necessarily restrains
interstate commerce. Moreover, it asserts that Pennsylvania’s
in-state processors are so entrenched (we noted previously that
“in-state dealers dominate the wholesale milk market in
Pennsylvania,” Cloverland 

), other means of
competition are so lacking, and the wholesale minimums are set
so high (such that they effectively set the actual price for all


   16
      Cloverland’s exemption from the over-order premium is
due to its status as an out-of-state firm, but it argues this fact is
not relevant to the heightened scrutiny analysis.

                                 32
handlers), that no out-of-state firm can compete in the
Pennsylvania market without offering a lower price. As we
summarized the argument, “[p]reventing dealers from attracting
customers by offering lower prices . . . helps in-state dealers
maintain their traditional hegemony.” 

a Competitive Advantage Arise from
                     Cloverland’s Status as an Out-of-State
                     Firm?




        Although Cloverland’s arguments were persuasive on the
limited facts (construed in Cloverland’s favor) available in
Cloverland I, they are less compelling in light of the factual
record developed at trial. As a starting point, we note that
Cloverland is incorrect that, at least on the facts of this case,
state regulations that nullify a competitive advantage unrelated
to the firm’s out-of-state status may nonetheless, on their own,
trigger heightened scrutiny.

        As the District Court framed the issue, “courts have
struggled with . . . the difficulty in examining laws which do not
facially discriminate against out-of-state interests under
[d]ormant Commerce Clause jurisprudence. When does a
simple economic advantage rise to the level of [a] competitive
advantage, within the heightened scrutiny context?” Cloverland
II, slip op. at 16-17. The Supreme Court has never expressly


                               33
addressed this issue, but it has often stated that a state law that
“cause[s] local goods to constitute a larger share, and goods with
an out-of-state source to constitute a smaller share, of the total
sales in the market” is unconstitutional “because it, like a tariff,
‘neutraliz[es] advantages belonging to the place of origin.’”
West Lynn 

(quoting 

) (internal quotation marks omitted) (emphasis
added); Polar Ice 

(same with respect to
a Florida law barring in-state handlers from buying raw milk
from out-of-state producers); see also Wash. State 

(noting that a North Carolina statute forbidding
importing apples with a non-USDA quality mark “has the effect
of stripping away from the Washington apple industry the
competitive and economic advantages it has earned for itself
though its expensive inspection and grading system”).

       We have observed that “[t]he Supreme Court’s opinions
in Baldwin and Washington State Apple show that if a state
regulation has the effect of protecting in-state businesses by
eliminating a competitive advantage possessed by their out-of-
state counterparts, heightened scrutiny applies. . . . [Here,] a
reasonable trier of fact could find that Pennsylvania’s minimum
wholesale prices eliminate a competitive advantage enjoyed by
out-of-state dealers like Cloverland.” Cloverland 

. By referencing a “competitive advantage enjoyed by
out-of-state dealers” not shared by in-state dealers, we presumed
the advantage would have something to do with the out-of-state
status of the advantaged firms.


                                34
         Cloverland argues, however, that it need not prove a cost
advantage linked to its place of origin. It contends that the mere
fact it has some efficiency advantage (for whatever reason) over
at least one in-state firm means it can sell milk for less, and
because Pennsylvania’s minimum wholesale prices nullify its
ability to do so, they render meaningless its efficiency advantage
and thus trigger heightened scrutiny.

        The ramifications of this argument are expansive.
Cloverland’s proposed rule would eviscerate longstanding
Supreme Court precedent that “price regulation [is not] an
impermissible burden upon commerce” where “the burden on
commerce is indirect and only incidental to the regulation of an
essentially local activity,” Polar Ice 

(citing Milk Control Bd. of Pa. v. Eisenberg Farm Prods., 

(1939), and Highland Farms Dairy v. Agnew, 

(1937)) — subject, of course, to the Pike balancing test.
See Cloverland 

-16. As the District Court
noted, in any state with minimum price regulations, there will
always be some out-of-state competitors (like Cloverland) that
have some advantages over at least one in-state firm, even
though those advantages may reflect differences in capacity,
efficiency, workers, etc., that have nothing to do with the firm’s
location (and indeed are shared with many in-state firms). See
Cloverland II, slip op. at 16-17, 19. Thus, under Cloverland’s
proposed rule, state milk price regulations will always be subject
to heightened scrutiny (and thus nearly per se invalid), despite
the fact we and the Supreme Court have clearly recognized this


                               35
is not always the case. If the neutralization of any advantage
possessed by any out-of-state firm were enough to trigger
heightened scrutiny, it is difficult to think of a state pricing
regulation that could survive.

       The elimination of a particular competitive advantage
belonging to the place of origin is not, of course, a determinative
inquiry in every case. For example, where a state statute
imposes an outright ban on competition by reserving all business
for an in-state firm or firms and explicitly barring interstate
commerce, see 

-93; Polar Ice 

, or targets out-of-state firms for special
burdens from which in-state firms are exempt, see West Lynn

; Wash. State 

, the problem is more blatant: the out-of-state firms are
subjected to overt discrimination or an outright ban on
competition. But Cloverland does not allege those barriers here.
Rather, it alleges more subtle discrimination, stemming from a
law and regulatory scheme that are not facially discriminatory.
Hence, a nexus between the advantage that has been neutralized
and the firm’s out-of-state status is necessary.

      We therefore hold that, if Cloverland is to succeed in
demonstrating that heightened scrutiny applies because its
competitive advantages have been neutralized by the
Pennsylvania Milk Marketing Law, it must establish that those
advantages arise by virtue of its out-of-state status.



                                36
             2.     Does Cloverland Have an Out-of-State
                    Competitive Advantage?

      The question becomes, then, whether the Pennsylvania
Milk Marketing Law operates to neutralize competitive
advantages belonging to Cloverland’s place of origin. The
record reveals only one such potential advantage vis-a-vis
Cloverland’s Pennsylvania competitors — its potential raw milk
cost advantage due to its exemption from the over-order
premium.

                    a.     Legal Errors in the District Court’s
                           Analysis

      We note at the outset two legal errors in the District
Court’s opinion that cannot form the basis for a determination
of whether Cloverland’s exemption from the over-order
premium should trigger heightened scrutiny.

                           (1)     Nexus Between the
                                   Challenged State Law and
                                   Cloverland’s Advantage

       We are unpersuaded by the District Court’s reasoning
that because Cloverland’s exemption arises from Pennsylvania
law, any cost advantage it has is “ephemeral” — the argument
being that if the minimum wholesale prices were invalidated (as
Cloverland desires), its purported raw milk cost advantage


                              37
would disappear because the over-order premium would no
longer be viable. We assume, without deciding, that if the
minimum wholesale prices themselves conferred an advantage
on Cloverland due to its out-of-state status, it could not rely on
that advantage to challenge those prices because invalidating the
prices would necessarily nullify the advantage (though, of
course, Cloverland would have no incentive to challenge the
prices in that event). Here, however, Cloverland’s purported
cost advantage arises from the operation of a different aspect of
the regulatory scheme — the over-order premium — that is not
challenged in this suit.

        The District Court concluded, based on the Board’s
evidence, that if the minimum wholesale prices were
invalidated, the over-order premium would fall. But this is not
a question susceptible to definitive proof by the Board’s experts.
It may well be that invalidating the minimum wholesale prices
would place pressure on the Commonwealth’s legislature to
repeal the over-order premium or extend it to out-of-state
purchasers (both of which would nullify Cloverland’s supposed
advantage). The cost disadvantage to Pennsylvania handlers of
paying the over-order premium and then competing against out-
of-state handlers without the protection of minimum wholesale
prices might make the over-order premium less feasible or
encourage them to purchase out-of-state milk to avoid the
premium. But we note that, under the current system,
Pennsylvania handlers already have an incentive to purchase
out-of-state milk to maximize their profit margins, yet the over-


                               38
order premium remains strong. Moreover, even assuming
Pennsylvania could constitutionally impose over-order
premiums on sales of milk to out-of-state purchasers (an issue
we do not decide here), doing so would presumably damage the
Commonwealth’s ability to export milk by essentially adding a
surcharge on exports. And the Board has gone to great lengths
(as explained below) to demonstrate several legal options
handlers use to sell milk at below minimum wholesale prices in
Pennsylvania, which suggests that invalidating the minimum
wholesale prices would not necessarily harm the handlers’
ability to pay the over-order premium to producers.

       If the minimum wholesale prices were invalidated, the
Pennsylvania legislature would be faced with these thorny
policy questions and would have to weigh the competing
considerations to determine whether the over-order premium
provisions of 31 Pa. Cons. Stat. § 700j remain viable. There is
simply no way to determine, at this stage, whether the over-
order premium would necessarily be repealed or extended to
out-of-state purchasers of Pennsylvania milk. At best, the
Board’s evidence and the District Court’s opinion establish that
such an event would be likely. We hold, though, that the
possible removal of Cloverland’s purported out-of-state
advantage if it is successful on this appeal is insufficient to
render that advantage irrelevant to our heightened scrutiny
analysis.

                            (2)    Availability of Non-Price


                              39
                                     Competition

       We also are unpersuaded by the District Court’s reliance
on the availability of non-price competition in the Pennsylvania
market. As we explain below, if Cloverland proves that its
exemption from the over-order premium confers a cost
advantage over Pennsylvania handlers that is neutralized by the
Commonwealth’s minimum wholesale prices, heightened
scrutiny applies. The fact that Cloverland might be able to
compete for some retailer accounts on non-price bases is
irrelevant, because even if there are other potential paths into the
Pennsylvania market that may allow an out-of-state handler
successfully to obtain some business, neutralizing an out-of-
state price advantage alone offends the dormant Commerce
Clause (especially since the record amply demonstrates that
price is — or, at least, would be if competition on price were
allowed — an important factor retailers consider in choosing a
supplier). See Wyoming v. Oklahoma, 

, 455
(1992) (holding that a state may not insulate part of its market
from out-of-state competition while leaving other parts open,
because this “measures only the extent of the discrimination; it
is of no relevance to the determination whether a State has
discriminated against interstate commerce”).17


   17
      Indeed, the Board’s argument (which the District Court
accepted) is similar conceptually to the one advanced
unsuccessfully by the State of Oklahoma in Wyoming. There,
Oklahoma argued that it “set[] aside only a ‘small portion’ of the

                                40
       If Cloverland proves its Maryland residency confers a
cost advantage it wishes to exploit in the Pennsylvania market,
but Pennsylvania law neutralizes that advantage (and thus
unlawfully shields in-state handlers from price competition),
heightened scrutiny will apply regardless of the existence of
other competitive means that Pennsylvania has not neutralized.
In demonstrating that heightened scrutiny should apply to a state
law, a plaintiff like Cloverland need only prove that its out-of-
state residency confers competitive advantages that are
neutralized by the state law under review, thus preventing
competition in the area in which the plaintiff enjoys an
advantage. The plaintiff need not prove it is prevented from
entering the market through competition on all possible bases.



Oklahoma coal market” for in-state coal producers, “without
placing an ‘overall burden’ on out-of-state coal producers doing
business in Oklahoma.” 

. Here, the
Board essentially contends there is no “overall burden” on
handlers (including out-of-state handlers) because they may
compete for business on non-price factors, even though at least
part of the market would surely be more receptive to Cloverland
if it were allowed to exploit its alleged out-of-state cost
advantage to offer a lower price on wholesale milk. The fact
that Cloverland may compete on bases in which it does not
enjoy an advantage (and on which it would, therefore, have
considerable difficulty displacing an incumbent handler from an
established account with a retailer, especially a retailer that is
primarily concerned with price) resolve the constitutional
problem.

                               41
                     b.     Factual Support for District Court’s
                            Decision

        Notwithstanding these errors in the District Court’s
analysis, we conclude it was not clearly erroneous for the Court
to find that Cloverland did not sustain its burden of proving an
actual raw milk cost advantage belonging to its place of origin.
Cloverland’s proof of its alleged cost advantage consisted
primarily of testimony from Lawrence Webster, Cloverland’s
general manager, and Robert Havemeyer, a management
consultant who testified as an expert in the costs of processing
and delivering milk. Webster testified that, by not adding the
over-order premium to the 65% of raw milk it purchases from
independent Pennsylvania farmers, Cloverland saves about five
cents per gallon, which would allow Cloverland to lower its
wholesale prices enough to gain a significant price advantage
over Pennsylvania competitors. Havemeyer testified that, based
on a survey of Cloverland’s costs during a 13-week period in the
fall of 1998 (when demand was at its peak), and updated in
2002, Cloverland’s variable costs (i.e., costs that vary with the
volume of production) allowed it to sell its milk in Pennsylvania
below wholesale minimum prices.

       Notably absent from Cloverland’s offer of proof,
however, was any comparison to actual Pennsylvania
competitors’ costs, despite the fact this was crucial to
demonstrating a cost advantage vis a vis those competitors. At
most, Cloverland proved it was capable of selling milk for less


                               42
than the wholesale minimum, but (as explained above) this is
not in itself sufficient to trigger heightened scrutiny.

        Carl Herbein, who testified for the Board as an expert in
milk cost accounting, conducted the comparison with
Pennsylvania handlers that Cloverland’s expert did not, and
concluded that Cloverland’s total costs rendered it less efficient
than three of the four sample handlers. Herbein explained, and
the District Court agreed, that Havemeyer’s calculations were
unreliable because they used an unreasonably small time frame,
relied only on Cloverland’s variable costs (rather than looking
at total costs),18 and considered a time when demand was at its


    18
          We observed in Cloverland I that, based solely on
Cloverland’s evidence, it appeared that “fixing prices based on
average total costs” — which includes variable costs and fixed
costs “such as equipment, office space, and other ‘overhead’
expenses” — “significantly increases dealers’ profits because it
will be in their economic interest to sell additional units of milk
at any price above their average variable costs, even if below
their total costs. Outside Pennsylvania, in contrast, milk dealers
generally offer prices based on their average variable 

. Herbein and Havemeyer presented conflicting
evidence on this point at trial, and the District Court agreed with
Herbein that an analysis of total costs is necessary to take into
account fully the costs of production. Since this conclusion is
supported by Herbein’s testimony, and Cloverland has not
shown that the testimony is demonstrably untrue, the District
Court’s conclusion is not clearly erroneous.

                                43
peak rather than a period representative of the full year.
Herbein, and other witnesses offered by the Board, also
explained that the cost advantage resulting from Cloverland’s
exemption from the over-order premium vis a vis a particular
Pennsylvania competitor may be mitigated by the fact that
Cloverland pays a slightly higher federal price for raw milk
because it is located in Maryland,19 and many Pennsylvania
handlers avoid paying minimum wholesale prices by entering
tolling agreements20 and taking advantage of allowable


   19
      Although Pennsylvania Areas 1 and 4 and the State of
Maryland are all under the Northeast federal order, there are
numerous price differentials that set various minimum prices
throughout the order. See, e.g., U.S. Dep’t of Agric., Northeast
States Class I Differentials, available at
http://www.fmmone.com/New_Zone_Diffs/
NEZoneDiffMap.pdf.
   20
       A tolling agreement is a state-approved service contract
whereby a handler contracts with a large retailer that is
vertically integrated (i.e., processes its own milk) to process,
package, and deliver milk to the retailer. Essentially, the
handler acts as a proxy for the retailer’s milk processing
operation, and thereby avoids charging minimum wholesale
prices.
        We have previously held that “out-of-state dealers’
ability to enter into tolling agreements [does not] meaningfully
mitigate[] the burden on interstate commerce,” because only
about one-third of wholesale milk sales in Pennsylvania are
made pursuant to tolling, and “[t]he dormant Commerce Clause

                              44
discounts for high volume sales, partial service (i.e., delivering
the milk to a retailer but not stocking it in the cooler or ordering
future inventory), and multi-store customers. None of this was
reflected in Cloverland’s offer of proof at trial, which relied on
a straightforward comparison between the price of raw milk
with the over-order premium and the price of milk without it.

       Cloverland argues strenuously that the Board’s data is
flawed. It contends, first, that the Board purposely selected
sample handlers with lower operating costs than Cloverland, and
argues specifically against the inclusion of “Dealer 4”,21 which


does not allow Pennsylvania to hamper out-of-state dealers in
two-thirds of its wholesale market on the ground that they may
compete freely in the remaining third.” Cloverland 

(citing 

). We adhere fully to
this holding. Thus, the Board’s evidence at trial, and argument
on appeal, that tolling is available to Cloverland as a means to
avoid minimum wholesale prices do not defeat Cloverland’s
dormant Commerce Clause challenge. But, as explained below,
the plaintiff in a case like this must prove an out-of-state
advantage over similarly situated in-state handlers, and in
conducting this comparison the plaintiff must take account of
the circumstances of the in-state handlers over which it claims
to have an advantage. Since Cloverland did not analyze
Pennsylvania handlers’ costs, the District Court did not clearly
err in relying on the Board’s comparison.
   21
      The relative costs of milk handlers are trade secrets, and
thus the sample dairies are referenced by pseudonyms.

                                45
is not located within a federal order area and moves less than
25% of its total Class I milk volume into that area, and thus is
only partially regulated by federal law. See 7 C.F.R. § 1001.7;
see also Sani-Dairy, Div. of Penn Traffic Co, Inc. v. Espy, 

, 412 (W.D. Pa. 1993) (same under earlier federal
order regulations). Herbein testified, and Cloverland does not
dispute, that a partially regulated handler like Dealer 4 need not
account to the Producer Settlement Fund, and is therefore able
to pay producers more for raw milk (which gives it a price
advantage over fully regulated handlers, and a corresponding
cost advantage when the market price of raw milk is higher than
the federal minimum). See 7 C.F.R. § 1000.76. Since
Cloverland is fully regulated federally, and must account to the
Producer Settlement Fund, it argues a comparison with the
partially regulated Dealer 4 is irrelevant.

       Cloverland’s argument misses the mark. Dealer 4 would
clearly be a competitor in the Pennsylvania market; although
Dealer 4 cannot sell more than 25% of its total milk volume into
a federal order area and remain partially regulated, it could still
compete with Cloverland for retail outlets in Areas 1 and 4 with
the milk it does move into the order area. Dealer 4 is not,
therefore, a wholly irrelevant comparison. Cloverland’s
argument is really addressed to whether Dealer 4 is a proper
representative for comparison. Yet Cloverland had every
opportunity to submit evidence of similarly situated
Pennsylvania handlers with comparatively higher costs, and it
failed to do so. As the Board was the only one to submit


                                46
evidence of comparative costs, the District Court did not clearly
err in relying on the Board’s sample.

       Second, Cloverland argues Herbein erroneously relied on
the “plant blend” cost of raw milk (i.e., the weighted average of
the costs of different classes of milk used by the plant, which is
paid to producers), rather than a segregated Class I price, in
conducting his analysis of comparative costs. This, according
to Cloverland, artificially lowered the costs of plants with less
Class I usage than Cloverland. The Board counters that the
plant blend method of calculating costs is the norm in the
industry and comports with generally accepted accounting
practices.

       Although Cloverland’s argument may have some merit,22


  22
     Herbein justified his reliance on the plant blend cost — as
does the Board in its arguments on appeal — by noting that
plant blend is an accepted method of pricing the cost of raw
milk, and reflects the way handlers actually operate (i.e., by
purchasing Grade A raw milk and only later deciding how it will
be classified). While we have no doubt plant blend is
recognized and accepted in the industry, here Cloverland
contends that its exemption from the over-order premium (which
applies only to purchases of Class I raw milk) gives it an
advantage over other purchasers of Class I raw milk. Reliance
on the plant blend cost in determining Cloverland’s advantage
over other handlers makes sense only if Cloverland and the
Pennsylvania handlers to which it is compared have the same

                               47
we note that it submitted no evidence that the sample handlers
(which Herbein testified were “very similar” to Cloverland in
terms of size, volume, and gross sales) had appreciably different
Class I usage. Havemeyer provided evidence of Cloverland’s
Class I costs, but did not compare these to any actual
Pennsylvania competitors, and the District Court expressly
found his study unreliable due to its limited time frame and
examination of only variable costs. Again, under these
circumstances, we cannot conclude that the District Court


Class I usage. Otherwise, as Cloverland argues, the comparison
will be flawed: because it has a high Class I usage, its plant
blend cost will be artificially higher than a plant with lower
Class I usage, even though the cost of its Class I milk alone may
be lower than the Pennsylvania competitor.
        Of course, it may be difficult to calculate a discrete
“Class I” cost because if (as the Board persuasively argued, and
the District Court found) it is appropriate to consider the total
costs of production — including fixed costs like capital
investment, facilities, equipment, and other overhead — in
deciding whether one handler has a cost or efficiency advantage
over another, it would be an arduous task to separate out the
portion of fixed costs attributable to the production of Class I
milk, as many fixed costs are directed at processing raw milk
regardless of its end use. It may be, then, that plant blend is an
unavoidable consequence of the way handlers actually work,
though we still think it most useful if the two handlers compared
use roughly the same amount of raw milk as Class I. But, as we
explain, in this case Cloverland’s failure of proof makes it
unnecessary to resolve this question.

                               48
committed clear error by relying on the only comparative
evidence before it.

                      *    *    *    *   *

        To be sure, we do not mean to suggest that a defendant
in a case such as this may defeat heightened scrutiny merely by
showing that the plaintiff does not have a competitive advantage
over some of its putative competitors. Such a rule would be at
odds with the well-established principle that “‘a statute may be
invalid if it favors only a single or finite set of businesses.’”
Cloverland 

(quoting Harvey & 

). But this does not relieve a plaintiff like Cloverland
of its burden of proving, as part of its prima facie case of
discrimination, that it has an actual competitive advantage over
some of its prospective competitors because it is an out-of-state
firm. Cloverland should have established that its exemption
from the over-order premium gave it a relevant cost advantage
over similarly situated Pennsylvania competitors that translated
into an ability to sell wholesale milk at a lower cost.23 Had it

  23
      Herbein’s testimony indicated that Cloverland had a cost
advantage over one of the four sample dairies offered by the
Board (which Herbein selected because they were similarly
situated to Cloverland), but Cloverland offered no proof that this
cost advantage arose because of its out-of-state status. In any
event, given the totality of the Board’s evidence, the District
Court did not clearly err in deciding (implicitly) that evidence of
a cost advantage over one similarly situated Pennsylvania

                                49
done so, the Board could (of course) have attempted to rebut this
proof by showing that Cloverland did not have an actual cost
advantage over its competitors, or its cost advantage was not
related to its out-of-state status. The District Court would then
have had to weigh the evidence. But Cloverland’s evidence did
not compare its costs to the costs of its competitors, and thus the
Court did not commit clear error when it held that Cloverland
failed to sustain its burden of proving an out-of-state
competitive advantage vis a vis its Pennsylvania competitors
that is neutralized by the Commonwealth’s mandatory minimum
wholesale prices. In this context, we do not have facts calling
for heightened scrutiny.

       C.      Pike Balancing

        Having concluded that Cloverland failed to prove the
applicability of heightened scrutiny, we proceed to the Pike
balancing test. 

. As noted, when a law
“effectuate[s] a legitimate local public interest, and its effects on
interstate commerce are only incidental,” the court must
determine whether “the burden imposed on such commerce is
clearly excessive in relation to the putative local benefits.” 

Court concluded that Pennsylvania’s mandatory
minimum wholesale milk prices survive this test.


handler, owing possibly to Cloverland’s exemption from the
over-order premium, was not sufficient to trigger heightened
scrutiny.

                                 50
        Cloverland argues that the “benefits” conferred by the
minimum wholesale prices are merely economic protectionism,
“the very evil the Commerce Clause seeks to prohibit.”
Appellant’s Br. at 40. The District Court disagreed, crediting
the Board’s witnesses who testified that minimum wholesale
prices make it easier for handlers to pay the over-order
premium. The over-order premium, in turn, helps small,
independent dairy farms remain profitable without joining
cooperatives, which fosters market diversity and prevents a
possible future rise in retail prices. Indeed, the District Court
was persuaded by expert testimony offered by the Board that the
abolition of minimum prices in California led to consolidation
of farms and handlers into a few dominant market participants,
resulting eventually in a sharp rise in consumer prices. The
District Court also credited testimony offered by the Board that
if small farms went out of business, “the agricultural
infrastructure built around the industry will also be negatively
affected, resulting in the loss of dealers, feed stores,
veterinarians, and other business[es] that support the dairy
industry.” Cloverland II, slip op. at 9, 26.

       Although Cloverland presented testimony tending to
establish the opposite — that the abolition of minimum
wholesale prices would have no adverse effect on the dairy
industry, which would continue to thrive as it does in other
states — the District Court was entitled to base its ultimate
factual conclusion on the Board’s evidence. “As we have
recognized, the clearly erroneous standard of review does not


                               51
permit an appellate court to substitute its findings for those of
the trial court. It allows only an assessment of whether there is
enough evidence on the record to support those findings. That
a different set of inferences could be drawn from the record is
not determinative. . . . Where there are two permissible views of
the evidence, the factfinder’s choice between them cannot be
clearly erroneous.” Scully v. US WATS, Inc., 

,
506 (3d Cir. 2001) (citations and internal quotation marks
omitted); see Anderson v. Bessemer City, 

, 573-74
(1985) (same).

       The next question, then, is whether the putative benefits
of the Milk Law are outweighed by a significant incidental
burden on commerce. Cloverland offered testimony that the
minimum prices prevented it from competing in the
Pennsylvania milk market, and thus the burdens on interstate
commerce were substantial. As noted, the Board offered
testimony that several other forms of competition exist in the
Pennsylvania market,24 and therefore the minimum wholesale


 24
     Although we held above that evidence of other competitive
outlets is not relevant to the question of whether heightened
scrutiny applies — because, as explained, the only relevant
inquiry in making that determination is whether an out-of-state
competitive advantage is neutralized by Pennsylvania law — the
presence of other forms of competition is one of many factors
that may be taken into account under the Pike balancing test,
where the inquiry concerns the magnitude of the incidental
burden on commerce and the countervailing benefits of the law.

                               52
prices do not entirely prevent new competitors from challenging
incumbent handlers and moving their milk interstate. As with
the benefits of the law, the District Court was forced to choose
between conflicting evidence, and we cannot say its choice was
clearly erroneous. Thus, we affirm the District Court’s holding
that Cloverland failed to sustain its burden of proving that the
incidental burdens of Pennsylvania’s minimum wholesale milk
prices outweigh the putative benefits.

                        V. Conclusion

       In Cloverland I, we expressed our unease with the fact
that Pennsylvania is the only state with mandatory price controls
of the sort described in this case, while its milk industry
(dominated by in-state firms) is flourishing. In our view, the
Board’s argument that the price controls are the only thing
preventing a vulnerable milk industry from complete collapse
(and a concomitant disruption in the supply of milk to
Pennsylvania’s residents) rings hollow in light of the undisputed
evidence that Pennsylvania exports far more milk than its
residents consume,25 and other states appear to do quite well




    25
       In Cloverland I, we noted that “Pennsylvania’s dairy
industry is among our nation’s most productive. Milk
production in the Commonwealth outpaces consumption by
roughly 350%. Annual production per-capita [for Pennsylvania
residents] is around 900 pounds; consumption [by Pennsylvania

                               53
without mandatory minimum prices. It is quite likely, as we
stated on the limited factual record in Cloverland I, that
preventing competition based on price works to the advantage
of Pennsylvania’s resident handlers by helping maintain their
dominance in the market. Indeed, there is apparently no
political incentive for in-state handlers to resist the price
controls, as evidenced by the fact that the Pennsylvania
Association of Milk Dealers has intervened in support of the
challenged law. Cf. West Lynn 

(“[O]ne would ordinarily have expected at least three groups to
lobby against the order premium, which, as a tax, raises the
price (and hence lowers demand) for milk: dairy farmers, milk
dealers, and consumers. But because the tax was coupled with
a subsidy, one of the most powerful of these groups,
Massachusetts dairy farmers, instead of exerting their influence
against the tax, were in fact its primary supporters. . . . [Thus,]
[the] State’s political processes can no longer be relied upon to
prevent legislative abuse.”).

       Cloverland failed to prove its case, however, and the
District Court cannot be faulted for relying on the evidence
presented at trial. But our opinion today does not settle the
question of the Pennsylvania Milk Marketing Law’s
constitutionality. Indeed, if another out-of-state plaintiff can
prove it has competitive advantages over actual Pennsylvania
competitors belonging to its place of origin (whether by virtue


residents] per-capita is merely 200 

.

                                54
of its exemption from the over-order premium or otherwise),26
and these advantages translate into an actual ability to sell milk
for less than similarly situated Pennsylvania handlers that is
neutralized by the minimum wholesale prices, heightened
scrutiny would apply.27 Cloverland lost this case because its
evidence was insufficient, but the constitutionality of
Pennsylvania’s minimum wholesale prices remains unresolved.

         In this context, the District Court’s judgment is affirmed.




   26
       The District Court noted, for example, that “if . . . [a]
competitive price advantage arose from [out-of-state] firms’
ability to buy cheap local milk, the wholesale price floor would
fall under a heightened scrutiny analysis.” Cloverland II, slip
op. at 18 n.13.
    27
        We have already held that “[i]f they are subject to
heightened scrutiny, the wholesale price floors cannot satisfy the
dormant Commerce Clause” because Pennsylvania could
“achieve its objective through alternative measures that do not
discriminate against interstate commerce” — by acting as a
market participant, for example. Cloverland 

.
This holding would, of course, apply in any future challenge to
the minimum wholesale prices.

                                 55