2.7....

. .£9 1... .
W414 as?
. .. .amwmws

It

55....“ . ,
a... 1.0-5.3. .. 0..
{I}? ina!‘ I

My»... .9»...
QR

. .s
. it. It... :1
. . . . a.i.-......s4r.?.£smmhd. .
. ‘ , I) Iv. :3“!
.4..- .2. x
.2 a... . .
.

and.

3' ;.¢.?'!?.

I.
S ﬁn I 31'
u >535
{'9an
.. i.
. a
..
. . .4."
i. .5»! .
.3: 3...” up»? n
2mihwﬁm... h... 331..."?
.. .i .. .
. i: 334.. .

A
<

x. u :12?)
Litiotiix
31(1):...

iﬁﬁqF: a...“

x
I .7: 1
0’. .It‘lvalut.‘ . 1'9". I IE
3.1 571.55??? {light
«3.... T..I.¢« .. .5... . . 354...
771.131.1413....Ir
.. 11'505’1 33.5?

 

. A! .(v HEHSIIL': ”.49
1:2. 4.13:0 .‘Nt..cn.
93!!!!

J“... I?! u-
1 . u ' It
2 (1.1.): 79.»: I.

i .
'0'. \‘\l \(i

99.6 .2131 H1;
3.3.0.35”th-
0 .3! I... 7.

! II...
$5.911}... nvx. .
I... ‘95]..93. .{6n‘ ..l4..~4
..l\{.t!ﬁ..1u.t. d...

 

 

 

THESIS
Z.

200:)

 

Date

0-7639

lilillllllllilllllllllllﬂl

3 1293 02043 5333
LIBRARY

Michigan State
University

 

 

 

 

This is to certify that the

thesis entitled

MARKETING AGENCIES-IN-COMMON IN THE
MICHIGAN MILK MARKET: A STUDY OF
THE PRODUCERS' EQUALIZATION COMMITTEE, 1980 -
presented by 1 9 9 8

Duane Alan Banderob

has been accepted towards fulﬁllment
of the requirements for

 

_M1_81__Jegree in Agricultural Economics

M 4%

Major professor

Z//§/00

 

MS U is an Afﬁrmative Action/Equal Opportunity Institution

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m Wm.“

 

MlCHl

MARKETING AGENCIES-IN-COMMON IN THE
MICHIGAN MILK MARKET: A STUDY OF THE PRODUCERS’
EQUALIZATION COMMITTEE, 1980 -— 1993
By

Duane Alan Banderob

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Agricultural Economics

2000

MICH

Price-ﬁxing
proxisions ofthe C
cooperatites to for
fluid-milk price abr
rm'nirnum price is I.
underSIanding ofth.
(“Emma Premium

cane] agreements, 5‘

HE S bargalmnu

ABSTRACT
MARKETING AGENCIES-IN-COMMON IN THE
MICHIGAN MILK MARKET: A STUDY OF THE PRODUCERS’
EQUALIZATION COMMITTEE, 1980 - 1998
By

Duane Alan Banderob

Price-ﬁxing by agricultural producers is allowed through cooperatives under the
provisions of the Capper-Volstead Act of 1922. Capper-Volstead also allows
cooperatives to form ‘marketing agencies-in-common’ that act as cartels and bargain for a
ﬂuid-milk price above the federally mandated minimum price. The difference above the
minimum price is known as an over-order premium. This research furthers the
understanding of the existence, operation, and stability of ﬂuid-milk cartels that generate
over-order premiums. The primary goal is to provide an economic analysis of ﬂuid-milk
cartel agreements, strategies, and stability.

This thesis concludes that over-order premiums exist in Michigan as a result of an
oligopolistic market structure. A formal cartel agreement exits because of the one
member’s bargaining power. Proprietary ﬁrms join the cartel only after a credible threat
to proﬁts is made. If this threat decays over time, independent farmers exit the cartel to
increase proﬁts. These conclusions assist in the ﬁIture evaluation of such associations, and

in the development and analysis of ﬂuid-milk cartel policy.

Copyright by
DUANE ALAN BANDEROB
2000

 

As with all \
involvement and co:
Agricultural Econo
well as funding an a
Larry Hamm‘s relatit
in making this thesis
depicted the complc
Producers Equalizz
“Emissary for this tl
research Christop'
and teacher cannot
knowledge and pzu
re“ﬂing draﬁs of

returned With Cour

ACIGQOWLEDGEMENTS

As with all work such as this, many people deserve recognition and thanks for their
involvement and contributions. Michigan State University and its Department of
Agricultural Economics top this list for providing an outstanding faculty and curriculum as
well as ﬁmding an assistantship which resulted in a rewarding educational experience.
Larry Hamm's relationship and rapport with Michigan dairy cooperatives was fundamental
in making this thesis a reality. Additionally, his guidance insured this research accurately
depicted the complexities and captured the history of the Michigan milk market. The
Producers' Equalization Committee and its member cooperatives provided essential data
necessary for this thesis, and more importantly recognized the future value of this
research. Christ0pher Wolf‘s involvement and contribution as major professor, mentor,
and teacher cannot be overstated. The quality of this thesis directly reﬂects his economic
knowledge and passion for sound and thorough research. Countless hours were spent
reviewing drafts of chapters, outlines, titles, tables, and ﬁgures; all of which were promptly
returned with countless green marks indicating potential areas of improvement. His
patience with this process certainly enhanced my learning and development (especially in
the area of writing), resulting in a much more valuable research experience. On the home
front, Cristin's never-ending support for this endeavor (including a move from California
to Michigan), helpﬁil edits, and partnership in a wonderful marriage were critical factors in
the success of this work, to which I owe an enormous thank you. Finally, the friendship of
Curtis and Liz, and arrival of Hannah Elizabeth have made East Lansing unforgettable. As

for my parents and their love and support, none of this could be possible without you.

iv

LIST OF TABLI
LIST OF FlGL'R

Chapter l: Introc
l.l Backgrou
1.2 Motivatic
1.3 Objective
1.4 C ontribut

Chapter 2: US. l
2.1 Early Mil

2.2 Dairy Ma
2.2.1 COOPCIaI
2.2.2 Enabling
2.3 History 0
3-4 Currenth
:41 Ennct l0n:
:43 Typesof
..4.3 Cuffenﬂ

Chapter 3: MiChi

ii Nona
3'; €0retic
3i] Modeler
3.4.2.3

bib Bars
-.. 3 cm

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................... vii
LIST OF FIGURES ..................................................................................................... viii
Chapter 1: Introduction .................................................................................................. l
1.1 Background ........................................................................................................ 2
1.2 Motivation .......................................................................................................... 3
1.3 Objectives ........................................................................................................... 4
1.4 Contributions ...................................................................................................... 6
Chapter 2: US. Milk Marketing Policies and Institutions ................................................ 7
2.1 Early Milk Marketing Policies and Institutions ..................................................... 8
2.2 Dairy Marketing Cooperatives ............................................................................. 9
2.2.1 Cooperative Classiﬁed Pricing and Revenue Pooling ......................................... 10
2.2.2 Enabling Legislation for Dairy Cooperatives ...................................................... 11
2.3 History of the Federal Milk Marketing Orders ................................................... 12
2.4 Current Milk Marketing Policies and Institutions ............................................... 13
2.4.1 Emotions of Dairy Marketing Cooperatives ....................................................... 14
2.4.2 Types of Milk Marketing Cooperatives .............................................................. 14
2.4.3 Current Operation of the Federal Milk Marketing Orders .................................. 16
Chapter 3: Michigan Milk Marketing ............................................................................ 18
3.1 History and Structure of Michigan Milk Marketing ............................................ 18
3.2 Theoretical Framework of Michigan Milk Marketing Organizations ................... 23
3.2.1 Model of Proprietary Firms ............................................................................... 27
3.2.2 Models of Milk Cooperatives ............................................................................ 28
3.2.2a The Bargaining Cooperative .............................................................................. 29
3.2.2b Balancing Cooperative ...................................................................................... 32
Chapter 4: Michigan Milk Cartels and Market Performance .......................................... 37
4.1 Establishment and Grth of the Producers’ Equalization Committee ............... 37
4.2 The PEC and the Michigan Milk Market ........................................................... 42

Chapter 5: Recent History of the PEC: The Price Wars ................................................. 46

5.1 PEC Price Wars ................................................................................................ 46
5.1.1 ThePrice War ofl982 ...................................................................................... 49
5.1.2 Thech War ofl992 ...................................................................................... 51
5.1.3 The Price War of 1997 ...................................................................................... 54
5.2 The PEC Agreement for Pooling and Transferring Over-Order Premiums .......... 55
5.2.1 The 1985 PEC Agreement ................................................................................. 56
5.2.2 The 1992 PEC Agreement ................................................................................. 60
5.2.3 The 1997 PEC Agreement ................................................................................. 65
Chapter 6: Cartels, Price Wars, and the PEC ................................................................. 72
6.1 Cartel Theory .................................................................................................... 72
6.1 .1 Cartel Conduct .................................................................................................. 73
6.1 .2 Cartel Enforcement ........................................................................................... 74
6.1.3 Monopoly Power .............................................................................................. 77
6.1.4 Cartel Theory and the PEC ................................................................................ 78
6.2 Price Wars ........................................................................................................ 81
6.2.1 Price Wars in Busts and Booms ......................................................................... 81
6.2.2 Equihbrium Price Wars ...................................................................................... 82
6.2.3 Bargaining Price Wars ....................................................................................... 84
6.2.4 Determinants of Price Wars ............................................................................... 85
6.2.5 Price Wars and the PEC .................................................................................... 87
Chapter 7: Non-Cooperative and Cooperative Games ................................................... 89
7.1 The Theory of Games ........................................................................................ 89
7.2 The Basics of Cooperative Games ..................................................................... 94
Chapter 8: Explaining Conduct in the Michigan Fluid Milk Market ............................. 101
8.1 Summary and Conclusions of Conduct in the Michigan Fluid Milk Market ....... 128
Chapter 9: Conclusions ............................................................................................... 131
9. 1 Contributions .................................................................................................. 13 1
9.2 Further Research ............................................................................................. 135
References .................................................................................................................. 138

Table 3.1

Table 3.2

Table 4.1
Table 5.1
Table 5.2
Table 5.3

Table 5.4

Table 7.1
Table 8.1
Table 8.2

Table 8.3

    

Statistic

PEC Cl.
C haracte
PEC Prit
Variable

PEC Firn
Agreeme

Basic Gar

 

COOperati

Table 3.1

Table 3.2

Table 4.1

Table 5.1

Table 5.2

Table 5.3

Table 5.4

Table 7.1

Table 8.1

Table 8.2

Table 8.3

LIST OF TABLES

Statistics for Southern Michigan Order 40 ................................................... 20
1998 Average Sales and Market Shares for Michigan Milk Marketing
Organizations .............................................................................................. 23
PEC Class I Premiums and Descriptive Statistics ......................................... 42
Characteristics of PEC Price Wars ............................................................... 48
PEC Pricing Schedule of 1985 ..................................................................... 51
Variables Used in PEC Pooling Mechanism Equations ................................. 56
PEC Firm Net Payoffs Per Hundredweight Calculated Under Different PEC
Agreements for May Quantities in December 1998 Terms ............................ 70
Basic Game Theory Terminology .............................................................. 100
Premiums Generated from Tacit Collusion in 1984 .................................... 113
Cooperative 1991 Monthly Averages in Million Pounds ............................. 120
Summary of Propositions and Conclusions ................................................ 130

vii

Figure 3.l Order
Figure 3.2 Order
Figure 4.1 Timeli
Figure 4.2 PEC C
Figure 4.3 The Pl
Figure 5.1 PEC F
Figure 5.2 PEC F
Figure 5.3 PEC F
Figure 5.4 PEC F
Figure 7.1 Emmr

Figure 3-1 Static ?
E111d Cc

Figure 8.2 Non-C

COOpe

LIST OF FIGURES

Figure 3.1 Order 40 Annual Milk Production and Utilization: May 1980 — Dec. 1998...21

Figure 3.2 Order 40 Weighted Annual Minimum Class Prices in Dec. 1998 Dollars ...... 22

Figure 4.1 Timeline of the Michigan Milk Cartels ......................................................... 39
Figure 4.2 PEC Class I Premiums from May 1980 through December 1998 .................. 41
Figure 4.3 The PEC Effect on Market Performance for May 1997 ................................ 43
Figure 5.1 PEC Firm Payoﬂ°s Under the 1985 Agreement in Dec. 1998 Terms .............. 60
Figm'e 5.2 PEC Firm Payoﬂ‘s Under the 1992 Agreement in Dec. 1998 Terms .............. 65
Figure 5.3 PEC Firm Payoﬁs Under the 1997 Agreement in Dec. 1998 Terms .............. 68
Figure 5.4 PEC Firm Payoﬁ‘s Under Actual Agreements in Dec. 1998 Terms ................ 69
Figure 7.1 Example Game Matrix ................................................................................. 90

Figure 8.1 Static Non-Cooperative Game In the Class I Market between Cooperative M
and Cooperative B ..................................................................................... 105

Figure 8.2 Non-Cooperative Game In the Class I Market between Cooperative M and
Cooperative B Including Balancing ............................................................ 1]]

viii

 

The C appe
cooperative bargai
not 'monopolize‘ c
amociations exist ll
more stable fann-n

The C appei
common‘ between
producers or those
agencies-in.comm0
fora milk Price abo
minimum Price or tl

existence of these p
for The Mket'widg
market conditions 5

This thesis f
ﬂuid-milk marketing
PM goal is to p]
common agrwmem:

Such 3880Ciations an

it‘lICt

Chapter 1: Introduction

The Capper-Volstead Act of 1922 extends limited exemption to farmers’
cooperative bargaining and marketing associations from anti-trust laws, so long as they do
not ‘monopolize’ or ‘restrain trade’ resulting in ‘undue price enhancement.’ Such
associations exist in the dairy industry to gain bargaining power resulting in higher and
more stable farm-milk prices.

The Capper-Volstead Act allows cooperatives to form ‘marketing agencies-in-
common’ between independent cooperatives, or cooperatives and independent dairy
producers or those producers not belonging to a cooperative association. Marketing
agencies-in-common coordinate pricing activities between participating ﬁrms and bargain
for a milk price above the federally-mandated minimum price. The difference above this
minimum price or the federal order price is known as an ‘over-order premium.’ The
existence of these premiums is attributed to cooperative market power and compensation
for the market-wide services cooperatives perform (Babb). These premiums also reﬂect
market conditions such as demand elasticities for raw milk.

This thesis furthers an understanding of the existence, operation, and stability of
ﬂuid-milk marketing agencies-in-common that generate over-order premiums. The
primary goal is to provide an economic analysis of ﬂuid-milk marketing agency-in-
common agreements, strategies, and stability. The results assist in the ﬁiture evaluation of

such associations and in the development and analysis of marketing agency-in—common

policy.

 

1,1 BacI-igl’oul

In 1956. M
premiums in the ﬁt
Michigan marketin.
redistributing the 0
Michigan ﬂuid-mill
Equalization C omn

The PEC in
consisted of mo m
collectively market
State while proprici
Pertaining 15 perce
b‘lkhigan milk pro<
Produced. This m,
on Class I milk. Or
on Class 1] milk 0:
represented appro)
receipts generate d

ECoanic 1

F .

IncEntives to (
Wes -

0f mCCntiVes

1.1 Background

In 1956, Michigan cooperatives initiated the practice of collecting over-order
premiums in the ﬂuid-milk market through a marketing agency-in-common. Since then,
Michigan marketing agencies-in—common have included provisions for pooling and
redistributing the collected premiums between participating members. Since 1966, the
Michigan ﬂuid-milk marketing agency-in-common has been known as the Producers’
Equalization Committee (PEC).

The PEC in 1998 covered milk marketed in Federal Milk Marketing Order 40 and
consisted of two major cooperatives and several proprietary ﬁrms. These cooperatives
collectively marketed approximately 85 percent of the total milk volume produced in the
state while proprietary ﬁrms that represented independent farmers controlled the
remaining 15 percent of the market’s milk. In 1996, the PEC returned to Grade A
Michigan milk producers an average 46 cent per hundredweight premium over all milk
produced. This average premium was the result of an 88 cent per hundredweight premium
on Class I milk, or milk used for ﬂuid products, and a 46 cent per hundredweight premium
on Class II milk, or milk used in soﬁ products such as yogurt. The premimns of 1996
represented approximately $20.2 million, or 2.5 percent of the $804 million total cash
receipts generated ﬁom milk sales in Order 40.

Economic theory suggests that cartel agreements are not stable because members
face incentives to deviate ﬁ'om the agreement to increase proﬁts. PEC members face these
types of incentives. Since 1980, the PEC has experienced three ‘price wars’ where at least
one participating ﬁrm deviated from the current PEC agreement. A price war occurs

when a PEC organization that handles ﬂuid milk believes a better outcome can be attained

 

outside the PEC. o

redistribution of pr

eventually re-grou
The PEC mt
proprietary mill; prc
independent farmer:
“'1'“ agreement in h
oithheld by Proprie
1.2 Motivation
AllhOugh Cc
Premium in m Ost L
understood. Availa
most contempOrar},
pith regard to fedm
commit. demand c

The PEC ha

Michigan ﬂuid-milk

whiciently inelastic

Cartel agreement ( P

tumble Cartels be

outside the PEC, or desires a change to the rules that govern the pooling and
redistribution of premiums between members. In each case of these price wars, the PEC
eventually re-grouped and continued operation under a new PEC agreement.

The PEC most recently experienced a price war in September 1997 when two
proprietary milk processing ﬁrms stopped pooling collected premiums from their
independent farmers on Class I and Class II milk. The PEC resumed operation under a
new agreement in May 1998, but only after foregoing $4.2 million in potential premiums
withheld by proprietary ﬁrms.

1.2 Motivation

Although cooperatives and marketing agencies-in-common collect over-order
premiums in most US. ﬂuid-milk markets, the existence of these premiums is not widely
understood. Available literature explains the level of observed over-order premiums, but
most contemporary work does not concisely explain the existence of over-order premiums
with regard to federal milk marketing orders, cooperatives, marketing agencies-in-
common, demand characteristics, and market structure.

The PEC has successﬁrlly collected over-order premiums for over 40 years in the
Michigan ﬂuid-milk market. Stable cartel agreements require that: market demand is
sufﬁciently inelastic, the cartel controls ample supply, and cartel members adhere to the
cartel agreement (Pindyck and Rubinfeld). Usually, this last requirement results in
unstable cartels because rational ﬁrms act out of self-interest as there is always an
incentive for cartel members to deviate ﬁ'om the agreement. Without enforceable
contracts or credible threats, member ﬁrms lower their price below the cartel price, or

expand output beyond its cartel allotment, to increase sales and proﬁts. Therefore,

 

economic theory 5
the federal order r
Additional
surrounding the CE
organization theor
(Scherer). Comer
booms. Other exp
demand ﬂuctuatior
instigate changes ir
(levenstein ).
1.3 Objectives
This researt
conduct. and stabili

Common The use t

economic theory suggests that the price for Class I and Class II milk should tend toward
the federal order minimum price.

Additionally, conﬂicting opinions and empirical evidence mark the literature
surrounding the cause of cartel price wars. For example, traditional industrial
organization theory suggests ﬁrms revert to competitive behavior in times of recession
(Scherer). Conversely, Rotemberg and Saloner model price wars as occurring during
booms. Other explanations include price wars that occur in response to unanticipated
demand ﬂuctuations (Green and Porter) and price wars being a method for ﬁmts to
instigate changes in the cartel agreement for a more favorable distribution of cartel rents
(Levenstein).

1.3 Objectives

This research has several objectives related to understanding the existence,
conduct, and stability of over-order premiums and ﬂuid-milk marketing agencies-in-
common. The use of a unique data set from the PEC assists in determining the stability
and the cause of price wars in these associations. The use of this kind of data is unusual
and is the ﬁrst of its kind ﬁ'om U.S. agricultural cartels, as this price and quantity
information is usually considered proprietary information. The objectives of this research
are:

1. Assess the role of cooperatives in milk marketing,

2. Assess conditions in the Michigan milk market that led to the creation of the PEC,

3. Review operation, mechanics, and performance of the PEC since 1980,

4. Model ﬁrm interaction and conduct in the Michigan milk market using a game

theory framework, and

 

5. Evalu

and ti“.

The ﬁrst 2
surrounding the F
milkmarlceting. l
of cooperatives. ti
cooperatives in th
context in which p
which cooperative
members is a nece
The third C
inter-workings anc
OVer~order pools ii
tho research eXpla
Shows the net effe‘
The foumh
ﬁnncondum in The
games With bOIh or
game them}, mOde
mgard the existenc

[it reCent PE C .

5. Evaluate stability and determine the conditions that contribute to price wars

and those that lead to stable agreements in the PEC.

The ﬁrst and second objectives entail examining the history and literature
surrounding the Federal Milk Marketing Orders and cooperatives in US. and Michigan
milk marketing. Particular attention is given to the conditions facilitating the development
of cooperatives, the Capper-Volstead Act, and the changing roles and strategies of
cooperatives in the dairy sub-sector. This history provides the economic background and
context in which producers and cooperatives act. Understanding the environment in
which cooperatives use bargaining power to generate over-order premiums for their
members is a necessary ﬁrst step to evaluating the existence and operation of the PEC.

The third objective utilizes the available data and literature to reveal the speciﬁc
inter-workings and history of the PEC. This thesis tracks the growth and development of
over-order pools in Michigan and examines the three price wars since 1980. Additionally,
this research explains the PEC’s pooling mechanism under each PEC agreement and
shows the net effects of each price war to participating ﬁrms.

The fourth and ﬁfth objectives require the application of game theory to model
ﬁrm conduct in the Michigan milk market. Speciﬁcally, cooperative and non-cooperative
games with both cooperatives and proprietary ﬁrms as the players are utilized. These
game theory models support the informal proofs of four propositions. These propositions
regard the existence of over-order premiums in Michigan; contributing causal factors of
the recent PEC price wars; and the interaction between independent farmers and

proprietary ﬁrms, and PEC cooperatives.

 

1.4 Contribut
This thesis
and marketing age]
addition it chroniC
the PEC since 1981
1980. and assesses
the analysis. this re
cooperative games.
theories of cartels a
This thesis 1
market structure be
rational However~
Pooling and redistri
i‘TWTIetries in shor
agreement and [hug
distribution Offents

of ‘
mCOmpatl‘ble dem

1.4 Contributions

This thesis combines the literature of milk marketing, cooperatives, federal orders,
and marketing agencies-in-common to explain the existence of over-order premiums. In
addition, it chronicles the history of Michigan marketing agencies-in-common focusing on
the PEC since 1980. This thesis compares and contrasts the four PEC agreements since
1980, and assesses the cause and effects of the three most recent price wars. To complete
the analysis, this research provides an overview of the theory of cooperative and non-
cooperative games, examines modem industrial organization theory and applies the
theories of cartels and price wars to the PEC.

This thesis ﬁnds that over-order premiums exist in Michigan as a ﬁmction of its
market structure because in the long-run tacit collusion between industry oligopolists is
rational. However, the PEC formalizes over-order premiums and includes provisions for
pooling and redistributing these rents among members. The PEC exists because of
asymmetries in short and long-run preferences between members. With a formal
agreement and thus an economic cartel, price wars occur as a means to renegotiate the
distribution of rents between members. Speciﬁcally, early PEC price wars occur because
of incompatible demands on the part of the PEC cooperatives. Independent farmers
rationally remain autonomous from the PEC and free-ride on its Class I and Class II
premituns, increasing proﬁts. However, with a credible policy threat to these proﬁts,
independent farmers, through proprietary ﬁrms, rationally join the PEC and pool their
premiums. If this threat fades over time, independent farmers exit the PEC and again free-
ride and increase proﬁts. Once the threat is restored, independent farmers rationally return

to the PEC.

 

C
Milk mar
extensive govern
witnessed the cre-
control the math
markets controlle
dominated urban
created cooperati
cooperatives were
desirable levels. tl
Milk has 5
iniTlllding the ﬂovv
Variation of milk ‘
han'CSted at least

Other day DUE It
COHVerted into a .

Supply and dema

Chapter 2: US Milk Marketing Policies and Institutions

Milk rmrkets throughout the world are complicated and almost always involve
extensive government policies and institutions. In the early 19003, the US. dairy industry
witnessed the creation of new policies and institutions to alleviate market disorder and to
control the marketing of milk and milk products. With many small farms supplying
markets controlled by few milk processors, large proprietary milk processing ﬁrms
dominated urban milk markets and farmer price negotiation. In response, producers
created cooperatives and lobbied for policies to counter this market power. Because
cooperatives were inadequate in stabilizing prices and increasing farmer income to
desirable levels, the Federal Milk Marketing Orders, or federal orders, were created.

Milk has several characteristics that contribute to a unique marketing system
including the ﬂow, perishability, and seasonal nature of milk production as well as the
variation of milk components used in manufacturing. Milk is a ﬂow product that is
harvested at least twice daily and is transported ﬁ'om the farm to market at least every
other day. Due to its perishable nature, milk must be moved to market quickly or
converted into a product with a longer shelf life. Milk has counter-seasonal peaks in
supply and demand as its production depends on the biological ﬁmction of dairy cows,
resulting in peak production in the spring and summer when fresh feed is readily available.
Conversely, demand for ﬂuid milk peaks in the fall and winter when school lunch
programs and the holiday season increase consumption. To satisfy ﬂuid dermnd in the fall
and winter with a ﬁxed number of cows, surplus milk is produced in the spring and
summer with peak production. Milk is a homogenous product at the farm level, but when

used in manufacturing, the retail values of the individual components vary. Mo st

cooperatives and

 

of a ditierenl raI i0
ditierenl Prices ant

Thése char
and marketing met
ﬁrms. an inherent l
of cooperatives an

This chaptt
cooperatives and I
early milk markets
cooperatives. Maj
classiﬁed pricing. r
chapter then focus:
llbtl'ument such as

the coth of toda

common

Early Milk

cooperatives and proprietary ﬁrms produce several products, each of which is composed
of a different ratio of fat, solids-not-fat, and ﬂuid carrier. Each of these products have
different prices and demand elasticities.

These characteristics of milk, its production and its processing, shape milk pricing
and marketing mechanisms. Without factors to offset the bargaining power of proprietary
ﬁrms, an inherent instability in milk marketing exists that has resulted in the involvement
of cooperatives and the federal orders.

This chapter reviews the history and current operation of US. dairy marketing
cooperatives and the federal orders. The chapter is organized chronologically, reviewing
early milk markets and the conditions leading to the formation of the ﬁrst dairy marketing
cooperatives. Major developments in cooperative milk marketing are assessed including
classiﬁed pricing, revenue pooling, and the laws allowing for producer cooperation. The
chapter then focuses on the inadequacies of the ﬁrst cooperatives and the need for a policy
instrument such as the federal orders. These institutions and policies are later reviewed in
the context of today’s marketplace with special attention given to marketing agencies-in-
common.

2.1 Early Milk Marketing Policies and Institutions

From the late nineteenth century through the early 19205, large proprietary ﬂuid-
milk processors dominated urban milk markets. Dairy farmers commonly sold milk to a
handﬁrl of proprietary ﬁrms in each market. Proprietary ﬁrms controlled price
negotiations and offered a uniform or ﬂat-milk price, regardless of end use, to all
producers. However, the price elasticities of the ﬁnished products allowed proprietary

ﬁrms and retailers to price discriminate with the dairy products they sold to consumers.

These PI
peak production
spring and sumrr
valued manufact
bearing the addit
ﬁrms balanced ll:
In some instance
manufacturing m
moved from urba
environment. dili:
2-2 Dairy M.-

A COOper
mWTShlP. contr
lBanOn‘). Memb
cOOllel’aliVQS ma:
ever)" agricuh Ufa
‘Cmppt.

Althoug}

The [Rt

These price plans found limited success and would break down in the months of
peak production (USDA, 1989). Proprietary ﬁrms paid a lower ﬂat-milk price during
spring and summer months for milk which they used in greater proportions for lower-
valued manufactured products. These prices often did not compensate producers for
bearing the additional cost of producing milk that was ﬂuid quality. Further, proprietary
ﬁrms balanced ﬂuctuating seasonal supplies by refusing to accept some producers’ milk.
In some instances, displaced producers could not ﬁnd an alternative milk market because
manufacturing milk plants, unable to meet the higher ﬂuid milk prices, had closed down or
moved from urban marketing areas (Manchester). To stabilize and improve this market
environment, dairy producers organized some of the ﬁrst agricultural cooperatives.

2.2 Dairy Marketing Cooperatives

A cooperative is a distinct business that operates under the three principles of user
ownership, control, and beneﬁts which is generally referred to as business-at-cost
(Barton). Members join cooperatives for primarily economic reasons, although
cooperatives may also pursue noneconomic objectives. Cooperatives operate in nearly
every agricultural business activity, but play their largest role in the US. dairy economy
(Crow)-

Although dairy cooperatives were created in the 18503 as local crearneries and
cheese plants, the ﬁrst cooperatives involved in ﬂuid-milk bargaining were organized in
the last quarter of the nineteenth century (USDA, 1995). These cooperatives focused

primarily on negotiating a higher ﬂat-milk price with the large proprietary ﬁrms.

22.1 COOP

After

quickly found
needed for flu
c00peratives i
Under

highly perishal
milk used in st:
and butter. rec.
Commonly em}-
lower price for
equal to the ma
Classmed PTiCinl
gTeater V'Olumeg
COOpera
classiﬁed PriCinL-
COOpemi‘e blen
for van'OuS Uses ‘
Mano” altd clz

C .
§

in .
Part because Pa

Home
(an.
5 ﬁmis (

2.2.1 Cooperative Classiﬁed Pricing and Revenue Pooling

After early cooperatives increased the ﬂat-milk price through bargaining, they
quickly found that increases in price led to a problem in disposing of surplus milk not
needed for ﬂuid uses (USDA, 1995). Recognizing differences in the value of milk,
cooperatives introduced classiﬁed pricing plans.

Under classiﬁed pricing systems, milk is priced based on its end use. In this way,
highly perishable ﬂuid milk with a very inelastic demand receives the highest price, while
milk used in storable manufactured products with a more elastic demand, such as cheese
and butter, receives a lower price (Wolf and Ham). The ﬁrst classiﬁed pricing plans
commonly employed three classes. The highest price was paid for milk for ﬂuid uses, a
lower price for ﬂuid cream, and the lowest price for milk used in manufactured products,
equal to the market price for manufacturing quality milk. Numerous cooperatives adopted
classiﬁed pricing plans in the early 1920s and 19303 as cooperatives gained control of
greater volumes of milk (USDA, 1995).

Cooperative revenue pooling plans were a natural complement to cooperative
classiﬁed pricing plans. With revenue pooling, cooperative members received a
cooperative blend price, or a weighted average of revenues from all cooperative milk sold
for various uses within a particular market. As different cooperatives had varying
utilization and classiﬁed prices, cooperative blend prices varied between cooperatives.
Cooperative classiﬁed pricing and pooling programs did not ﬁnd complete success at least
in part because participation, as well as cooperative membership, was voluntary.
Proprietary ﬁrms continued to purchase milk from independent famrers. The proprietary

ﬁrms that purchased milk from these farms usually had high ﬂuid utilization, but paid a

10

 

ﬂapmﬂk price
mill; price buy
competitors. \
able to cut ret;
vtith the ﬂat-rt
decrease in the
buyer—having
independent pr
Eventually all r
belted the po“
milk utilization
2.2.2 Enablir
AS coo;

and 1922, the C

ﬂat-milk price for ﬂuid milk comparable to the cooperative blend price. Since these ﬂat-
milk price buyers were purchasing milk for ﬂuid uses for less than their ﬂuid milk
competitors, who paid the higher cooperative classiﬁed price, ﬂat-milk price buyers were
able to cut retail prices to increase sales. In order for cooperatives to compete effectively
with the ﬂat-milk price buyer, the ﬂuid class price was decreased which resulted in a
decrease in the cooperative blend price. This process continued, as the ﬂat-milk price
buyer—having an advantage in price negotiation—lowered the ﬂat-milk price to
independent producers causing the cooperative to again lower their classiﬁed prices.
Eventually all milk received the same price regardless of use. Furthermore, cooperatives
lacked the power to audit the records of proprietary ﬁrms to assure accurate reporting of
milk utilization, resulting in widespread underpayment to cooperatives (Manchester).
2.2.2 Enabling Legislation for Dairy Cooperatives

As cooperatives gained prominence in marketing milk in the early 19003, these
producer-owned associations relied on federal legislation for their legal existence. In 1914
and 1922, the Clayton Act and the Capper-Volstead Act provided farmers the legal right
to form associations to collectively act on farmers’ behalf with limited exemption ﬁ'om
anti-trust laws. Capper-Volstead gives permission to dairy producers, and other
agricultural producers, to act together to set prices and collectively process, prepare,
handle and market their products within certain bounds.

Case law and legal precedence prevents cooperatives from engaging in coercion,
predatory practices, or price ﬁxing with third parties. Section 2 of Capper-Volstead
speciﬁcally prohibits cooperatives ﬁ'om acts that monopolize or restrain trade——as

interpreted by the US. Secretary of Agriculture—resulting in undue or unreasonable price

11

 

 

enhancement 0
enhanccmﬂll- :
inquiry by the I
no action.

The Ian
in-comrnon so '
marketing agen
belonging to a t
authorization ft
Chapter.

2-3 History

AS descr
increase farther
Plans because p]
Depression. 3 \v
exacerbated the
b." four fundame
depressed farme]

reduced consum
t

enhancement of their products. To date, seven cases of possible undue price
enhancement, six involving over-order premiums in milk markets, have been subject to
inquiry by the Department of Agriculture (Manchester). In all cases, the Secretary took
no action.

The language of Capper-Volstead also allows the formation of marketing agencies—
in-common so long as all members are agricultural producers. Cooperatives can operate a
marketing agency-in-common with other cooperatives, or with independent producers, not
belonging to a cooperative. This part of the Capper-Volstead Act is the legal
authorization for the PEC and similar associations, and is discussed in detail later in this
chapter.

2.3 History of the Federal Milk Marketing Orders

As described, early dairy cooperatives were unsuccessful in long-run attempts to
increase ﬁrmer income or sustain orderly marketing through classiﬁed pricing and pooling
plans because proprietary ﬁrms could break cooperative bargaining. During the Great
Depression, a weak national economy, large milk surpluses, and depressed ﬁrm prices
exacerbated the already existing problems in marketing milk. This era was characterized
by four fundamental problems in milk marketing: the existence of unstable markets,
depressed ﬁrmer prices, inadequate supplies of a pure and wholesome product, and
reduced consumer purchases (USDA, 1989).

The ﬁrst legislative actions aimed to resolve the crisis in the dairy industry were
the Agriculture Adjustment Acts of 1933 and 1935. These acts created marketing orders,

or marketing agreements between producers, processors, and retailers (USDA, I989).

12

After some legal difﬁculty, the Agriculture Marketing Agreement Act (AMAA) of
1937 established a permanent role for the federal government in the regulation of the dairy
industry. The AMAA concentrated on milk pricing and long-term market stability through
producer requested and approved regional milk marketing orders (USDA, 1989). Federal
Milk Marketing Orders require that all milk processors pay uniform, minimum class prices
for milk based on end use. Federal orders also ensure that market revenues are pooled
and divided between dairy producers through market-wide or individual handler pools. In
market-wide pools, total market revenues are divided across all milk and all producers
receive a weighted-average price. In individual handler pools, the receipts of an individual
handler or processor are divided across the milk delivered to that handler, and producers
are paid based on that handler’s utilization.

Through market-wide pooling the federal orders set marginal revenue equal to
average revenue for dairy producers. This occurs as an individual’s output does not affect
the blend price. Speciﬁcally, each producer supplies a negligible part of total output, thus
producers ﬁce only the market-wide pool price and perfectly elastic demand (Sumner and
Wolf).

2.4 Current Milk Marketing Policies and Institutions

Dairy marketing cooperatives and the federal orders, established in the early 19005
to improve the marketing of milk, remain the policies and institutions with the greatest
role in the current dairy economy. In 1998, more than 200 cooperatives controlled
approximately 86 percent of the nation’s milk supply. In addition, 68 percent of US. total
production is regulated by one of 30 existing federal orders (USDA Agricultural

Statistics).

13

2.4.1 Functio
To guar‘
are the primary
receive other he
payment. and pt
Dairy in
entire market. L
Spring peak pro
often shift the F
high costs int-0
CImperatives I]
remain idle wh.
and other marl.
rider PrOblem r
PIEijmS
2.4.2 Tb'pes
The be
deﬁne the Org:
dimintguisheS I
cOtipeﬁitives z

c .
CInCEntraIe 0‘

as .. .
.smng m

f uid u . .
tulzatio

2.4.1 Functions of Dairy Marketing Cooperatives

To guarantee members a market for their milk and to secure the best price possible
are the primary functions of today’s dairy marketing cooperatives. Cooperative members
receive other beneﬁts including representation in public policy issues, insurance of
payment, and personal satisfaction or utility in belonging to such an organization.

Dairy marketing cooperatives also perform balancing services that beneﬁt the
entire market. Market balancing is the changing of excess ﬂuid-milk production during
spring peak production into manufactured products to stabilize prices. Proprietary ﬁrms
often shift the handling and transportation of these reserves to cooperatives due to the
high costs involved and economies of scale in manufacturing dairy products (Manchester).
Cooperatives maintain excess capacity in processing plants to handle these reserves which
remain idle when ﬂuid market supply and demand are at equihbrium Market balancing
and other market-wide services are paid for by cooperative members and create a free-
rider problem within the industry if cooperatives are not compensated through over-order
premiums.

2.4.2 Types of Milk Marketing Cooperatives

The beneﬁts cooperatives provide, as well as the enterprises cooperatives choose,
deﬁne the organizational structure of dairy marketing cooperatives. This research
distinguishes between two forms of dairy marketing cooperatives, “bargaining”
cooperatives and “balancing” or full-service cooperatives. Bargaining cooperatives
concentrate on collective bargaining functions. They refrain from manufacturing, ﬁnding
bargaining power in the volume of milk controlled and seek proﬁts by attaining a high

ﬂuid utilization.

l4

Balancing or “manufacturing” cooperatives perform the ﬁmctions of bargaining
cooperatives in addition to converting market ﬂuid reserves into manuﬁctured products.
Balancing cooperatives improve bargaining power by using processing facilities to convert
reserves into storable hard products such as butter, powder, and cheese when market
conditions dictate.

Both types of cooperatives strive to maximize returns to members, but choose
different strategies and organizational structures to operate in diverse economic
environments and satisfy diverse producer expectations. It is important to note these
organizational and ﬁrnctional characteristics separate the market roles of dairy
cooperatives for economic analysis in later chapters.

To increase and strengthen market power, dairy producers and their cooperatives
can form marketing agencies-in—common. A marketing agency-in-common is a type of
cooperative including a group or combination of marketing cooperatives, or a group or
combination of marketing cooperatives and independent ﬁrmers who market products
under a common agreement. Marketing agencies-in—common assist dairy markets by
reducing marketing costs, disposing of surpluses in an orderly ﬁshion, and reducing
product price volatility (Cropp).

The PEC is a marketing agency-in-common currently operated by balancing and
bargaining cooperatives, with voluntary participation by Michigan’s independent dairy
ﬁrmers, some of whom sell their milk directly to proprietary ﬁrms. Proprietary ﬁrms
contract and secure a major portion of their raw milk supply through independent dairy

ﬁrmers who do not belong to a cooperative. Michigan proprietary ﬁrms never formally

15

join the PEC but purchase selected services ﬁom the PEC. The PEC is an organization of
raw-milk marketing ﬁrms.
2.4.3 Current Operation of the Federal Milk Marketing Orders

Through the federal orders allowance of classiﬁed pricing, price discrimination is
transferred to the ﬁrm level. With market-wide minimum class prices, marketing orders
essentially divide the market into at least three groups or classes, all with diﬂ‘erent demand
curves and varying demand elasticities for homogeneous milk. Michigan dairy producers
are paid for milk in four classes. Class I milk includes milk used for ﬂuid and beverage
purposes including pourable creams. Class II milk includes milk used in “soft” dairy
products such as sour cream, yogurt, and ice cream. Class IIIA milk includes milk used
for butter and nonfat dry milk powder, and Class III includes milk used for hard cheese.

Because the federal orders were also intended to increase ﬁrmer bargaining
power, dairy marketing cooperatives are given unique privileges in the structure of the
federal orders. Cooperatives are allowed to block vote for their members on matters of
order policy. As changes to federal orders require approval by one-half the producers and
two-thirds the production affected by a policy change, or two-thirds the producers and
one-half the production, large cooperatives are especially powerful in order politics.
Additionally, cooperatives can blend members’ revenues to pay prices below federal class
price minimums. However, this practice is limited as cooperatives must pay a competitive
price to retain membership. Finally, cooperatives may direct and pool milk in ways not
allowed proprietary handlers. Cooperatives can determine the best situation to increase
member returns by pooling across federal order boundaries. These privileges exist to

allow cooperatives to better serve the market.

16

More information about the history and operation of the federal orders and

cooperatives may be found in Manchester, and Bailey.

17

Chapter 3: Michigan Milk Marketing

In the early 19005, Michigan dairy farmers faced a market where milk processors
dominated price negotiations with individual farmers. To counter this market power,
dairy marketing cooperatives were organized throughout Michigan in the ﬁrst quarter of
the twentieth century. Cooperatives introduced classiﬁed pricing and pooling plans with
limited success. To introduce market-wide classiﬁed pricing and market-wide pooling,
dairy farmers adopted the ﬁrst Michigan federal order for the area surrounding Detroit in
1951. Since that time, the federal orders and dairy marketing cooperatives have guided
milk marketing in Michigan.

This chapter reviews the history of dairy marketing cooperatives and the federal
orders in Michigan. It also models the three ﬁrm types that participate in the Michigan
milk market: proprietary ﬁrms, bargaining cooperatives, and balancing cooperatives. The
ﬁrst section follows the development and grth of the cooperative movement in the
Michigan dairy industry, reviews the conditions leading to the creation of Michigan
federal orders and summarizes changes in industry structure since the 19505. The second
section presents ﬁrm optimization problems that are applied in later chapters to
understand behavior in the Michigan milk market.

3.1 History and Structure of Michigan Milk Marketing

As mentioned above, bargaining associations and marketing cooperatives gained
popularity in Michigan in response to processor dominated price negotiations. By 1946,
cooperatives were successful in raising farm milk prices, or gaining market power, in
eleven of the twelve major Michigan milk markets. Exceptions included producers in the

Lansing area, the northern Lower Peninsula, and the Upper Peninsula. In these markets,

18

cooperative organizations did not gain enough members or control sufﬁcient milk volume
to exercise market power in bargaining with proprietary milk processors.

In most major markets in which cooperatives operated, classiﬁed pricing and
cooperative pooling programs were established prior to the implementation of Michigan
federal orders in the 19505. Local sales committees participated in price bargaining and
were ultimately responsible for marketing milk in each local market.

In 1950, over 16,000 Michigan dairy farms supplied raw milk to 788 dairy
processing plants (Juevs and McMillan). These processing plants distributed ﬂuid-milk
and supplied manuﬁctured dairy products to residents of the twelve major metropolitan
markets.

In 1951, producers in the Detroit market approved the creation of the ﬁrst
Michigan federal order. Muskegon, Upstate (northern Lower Peninsula), and Upper
Peninsula orders followed later in the 19505. In 1960, as greater quantities of surplus
milk had ﬂowed into the Detroit market ﬁ'om unregulated areas, the Detroit order
expanded to include the southern Michigan region, excluding the Muskegon Order. In
1965, the Muskegon Order merged into the Detroit/Southem Michigan Order, and in
1973 the Upstate Order merged into the Southern Michigan Order, Order 40. The
geographic area covered by Order 40, the entire Lower Peninsula of Michigan, has
remained unchanged since 1973.

From this point forward, this research uses Order 40 and the Michigan milk
market interchangeably. Milk produced in the Upper Peninsula is regulated under a

separate federal order, Order 44, and does not inﬂuence the operation of the PEC. In

19

1993, productio
Order 40 and 0

These e:
fundamental ch.
l970s. as isolatt
began to align a

Table 3.
period the numl
more productivt
preferences alsc
dcsplle a growii

Table 3.1 Stati
\

, Numb.
\
19"3 7.77
1975 7 ,,
1980 6‘”:
.4
1985 5 9:.
1990 4:89
1995 3 74
1998 3’“
W
Figure 3

1998, production in the Upper Peninsula accounted for less than 2 percent of combined
Order 40 and Order 44 milk sales (USDA Market Administrator’s Report).

These expansions and consolidations of Michigan federal orders indicate
fundamental changes in the structure of the Michigan dairy industry. In the 19605 and
19705, as isolated urban milk markets expanded into one state market, local cooperatives
began to align and consolidate operations.

Table 3.1 summarizes the structure of Order 40 ﬁom 1973 to 1998. During this
period the number of producers steadily declined. However, farms became larger and
more productive as total deliveries remained relatively constant. Consumer tastes and

preferences also changed as ﬂuid-milk consumption and Class I utilization decreased

 

 

despite a growing population.
Table 3.1 Statistics for Southern Michigan Order 40
Number of Total Deliveries Class I Utilization Class I

Year Producers (million pounds) (million pounds) Utilization
1973 7,775 3,774 2,470 65%
1975 7,239 3,804 2,359 62%
1980 6,432 4,455 2,245 50%
1985 5,996 4,970 2,074 42%
1990 4,897 4,689 2,046 42%
1995 3,749 4,642 2,072 45%
1998 3,529 4,325 2,093 48%

 

Source: USDA Agricultural Statistics

Figure 3.1 shows the total production and utilization of milk supplied in Order 40

from May 1980 through December 1998. As the ﬁgure illustrates, total production and

Class I utilization have remained relatively constant, while Class II and Class III

utilization moved in opposite directions. Class II utilization increased beginning in 1985,

reaching a plateau in 1993. Class III utilization decreased beginning in the same year,

and reached a plateau in 1995.

Figure 3.1 OFdC

 

 

Million Pounds:
~.)
‘J
JO

Figure 3.:
31? inﬂated to De,
Dairy PTOduCts“ C
Price peaks in 198
The Class I price
Order 40 blend Pr
lowin1996 at $1.

The blend pricC gr;

 

Figure 3.] Order 40 Annual Milk Production and Utilization: May 1980 — Dec. 1998

 

 

 

 

 

 

 

450 l
as
g 300 -
a?
g 225 T . “(ﬂ/n-“Nswﬁnww-‘s‘MmN‘
3:3 1"”; ---------- thief. 5.1-«"6 “a‘ u"‘."°":"" ---------
:=: 1504 "" \' ~~~~~~
2 ﬁ“x.._________
75 s ------------------
o . . . . . , . . s , . .
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998
Year
T Total Prod - - - -Classl ....... Class 11 __..___. Class 111]

 

 

Figure 3.2 illustrates the minimum class prices for milk in the same period. Prices
are inﬂated to December 1998 dollars using the “Index of Prices Received by Farmers for
Dairy Products” compiled by the United States Department of Agriculture. The Class I
price peaks in 1990 at $19.48 and reaches its low in 1998 at $18.01 per hundredweight.
The Class I price increases over time until 1994 when it begins a steady decline. The
Order 40 blend price, the price farmers receive, peaks in 1994 at $18.02 and reaches its
low in 1996 at $17.15 per hundredweight. Except for the period between 1992 and 1994,

the blend price gradually decreases over time.

2]

 

20.00

19.00
E
3
. 631800
i 9 _
.2
L-
a.
17.00 .
16.00 -
198
\H,
In Decem

percent belongin;
cooperative and t
to either smaller 1
PTOpﬁeIM' milk 3

Combinec

production in Ore

Percent, and the t

PYOduction in

Ulllll' pTOdUCtion

Figure 3.2 Order 40 Weighted Annual Minimum Class Prices in Dec. 1998 Dollars

 

 

 

 

 

 

 

 

 

20.00 -
19.00 a .. .0. .0... a" ‘2 .\....... “o.
g . ......... -.' ..‘... g“ .0 .‘ .. 0“.
Q .
9 18.00 71
0
23’
17.00 4
16-00 I I T I T r T f 7 I I T MI I “~T T—-_‘—T-_—r_——_T—.~l
1980 1982 1934 1986 1988 1990 1992 1994 1996 1998
Year
....... ClassI - . . . Class 11 Class III 131ch

 

 

 

In December 1998, there were over 3,000 Michigan dairy ﬁrmers with about 70
percent belonging to two major cooperative milk marketing organizations, one bargaining
cooperative and one balancing cooperative. Remaining Michigan dairy farmers belonged
to either smaller bargaining cooperatives or contracted their milk to one of several
proprietary milk processors.

Combined, Michigan cooperatives market almost 85 percent of total milk
production in Order 40. In 1998, the balancing cooperative marketed approximately 65
percent, and the bargaining cooperatives jointly marketed 19 percent of total milk
production in Order 40. Proprietary ﬁrms marketed the remaining 16 percent of total
milk production.

However, these proportions are different for Class I utilization. In 1998,
proprietary ﬁrms controlled approximately 29 percent of Class I milk, the bargaining

cooperatives 27 percent, and the balancing cooperative the remaining 44 percent. Table

22

3.2 reports avera
individual ﬁrm. 1
sales in each clas

Table 3.2 1998 a
0'32
F
P
l Proprietary
¥ Firmsl 4‘3
l Bargaining;—
lCoo eratives: 4‘5
; Balancing"

Coo erativef 7b

‘Tital Market

 

 

I73

1
A

 

Order 40
historically prod
dairy Products. 1
rentesems aDPro:

P€rcent is the am

32 Theoretit
With the .
QCh pal-IICipat in.E

bChat-im in a co
n

details ofthe m
at.

3.2 reports average quantities and percentages by type of organization, rather than by
individual ﬁrm, for total pounds of milk sold, milk sales by class, and market share of
sales in each classiﬁcation for 1998.

Table 3.2 1998 Average Sales and Market Shares for Michigan Milk Marketing
0 anizations

Class I Class II Class 111 Total Market
Pounds Share Pounds Share Pounds Share Pounds Share

 

 

 

 

 

Proprietary
Firms 49,915,612 29% 1,263,985 2% 4,275,444 4% 55,455,042 16%

Bargaining
Cooperativﬁ 46,806,546 27% 14,745,333 19% 6,376,631 6% 67,928,510 19%

Balancing

Cooperative 76,865,881 44% 62,563,287 80% 93,545,920 90% 232,975,088 65%

 

Total Market 173,588,881 49% 78,572,605 22% 104,197,995 29% 356,358,639

 

 

 

 

 

 

 

Order 40 is a balanced market, or a market in equilibrium, as Michigan farmers
historically produce sufﬁcient milk to supply State demand for ﬂuid and manufactured
dairy products. Milk used in Order 40 reported monthly by the Market Administrator
represents approximately 95 percent of Order 40 total milk production. The remaining 5
percent is the amount of milk exported.

3.2 Theoretical Framework of Michigan Milk Marketing Organizations

With the market size, trends, and structure in mind, this section turns to modeling
each participating ﬁrm in the Michigan milk market assuming proﬁt-maximizing
behavior in a competitive environment. Prior to this, this section reviews operational
details of the market necessary to understand the ﬁrm models.

To simplify this analysis, the ﬁve market cooperatives in Michigan are modeled
as two organizations, one balancing cooperative and one bargaining cooperative. To

accomplish this, the Operations of the four bargaining cooperatives are combined into one

23

organization. Sl:
Therefore. forth
proprietary ﬁrm \

Milk prici

 

understanding is 1
price. officially tl
month. For exam
January 5. 1999.
Changes in value
minimum Class
BFP. The differ
Processors adva
minimum Class
November 5, 19
milk We’re 317:.
The ﬁrm
marketing Orgar
by the federal 0]
are UniqUe 10 Or
minimum Class
L‘ﬁe Price for the
Pmduqs Sold on

.or Class III mill

organization. Similarly, all proprietary ﬁrms are modeled as a single aggregate ﬁrm.
Therefore, for this research a balancing cooperative, a bargaining cooperative, and a
proprietary ﬁrm control all milk produced in the Michigan milk market.

Milk pricing in the federal orders is a complex process. However, only a basic
understanding is needed for this research. The monthly national minimum Class 111
price, ofﬁcially the Basic Formula Price (BFP), is announced on the ﬁfth day of the next
month. For example, the December 1998 Class IH price (BF P) was announced on
January 5, 1999. The BFP is announced with a one-month lag because it is based on
changes in value of butter and cheese traded on national commodity markets. Monthly
minimum Class I and Class 11 prices are calculated by adding a ﬁxed differential to the
BFP. The differentials are added to the BFP from two months prior to allow milk
processors advanced notice of the Class I and Class 11 price. That is, the December 1998
minimum Class I price was based on the October 1998 BFP price that was announced on
November 5, 1998. In December 1998, the Order 40 differentials for Class I and Class II
milk were $1.75 and $0.30 per hundredweight.

The ﬁrm models include the sale of all classes of milk by the representative milk
marketing organizations. Minimum prices for all classes are announced and monitored
by the federal order market administrator. Monthly minimum Class I and Class 11 prices
are unique to Order 40 due to the ﬁxed class differentials. However, the BFP and
minimum Class IH price is a national price for milk used in manufacturing and represents
the price for the residual milk supply used in producing storable, commodity dairy
products sold on the national market. Due to the national nature of this market, the price

for Class III milk is modeled as perfectly elastic.

24

Since the inception of the federal orders, milk uses and classiﬁcations have
changed. Most recently, several federal orders, including Order 40, split Class III milk
into two classes, Class III and Class IIIA. Before the split, the BFP was for milk used in
the production of cheese, butter, and dry milk powder. With the inception of Class IIIA,
milk used for cheese is priced at the BF P, while milk used for butter and nonfat dry milk
powder is priced at the Class IIIA price. Since all federal orders did not split Class III
milk, the national price announced for milk used in manufacturing, or BFP, applies to
Class III milk in all orders. However, these different Class III prices do not affect the
operation of the PEC, as the PEC uses only federal order Class I and Class 11 prices. The
PEC does not distinguish between milk sold for Class III and Class IIIA uses in the

monthly calculations.
Unique to cooperatives is a cooperative blend price, PEC, , which cooperatives use

to transfer proﬁts to farmer members. The cooperative blend price for cooperative C, is

the weighted average of milk sales by class, or:

PBCI : PICQ; + 1’17er + PIIIQICII

3.1)
( or +9; +92,

 

where Pf and P; are the cooperative class prices for milk sold for use in Class I and
Class 11, P1,, is the national or minimum price for milk sold for use in Class III, and

Qf, Qf, , and Qf” are the quantities sold by cooperative C in each class. Cooperative

prices for Class I and Class II milk include over-order premiums, a premium above the

federal order minimum price, thus are often higher than the minimum class prices. The

blend price is a ﬁmction of each cooperative’s prices and quantities sold. The formula for

calculating the blend price is the same for balancing and bargaining cooperatives, but the

25

actual blend price is different across cooperatives due to differences in class utilization
and actual class prices. That is, the potential exists for cooperatives to individually
collect over-order, quality, or volume premiums on the sale of raw milk. These
premiums may vary between cooperatives yielding different cooperative class prices.

In addition to balancing and bargaining, cooperatives provide a variety of other
member services that are not modeled in this research. For example, the beneﬁts and
costs from a guaranteed market, social capital, or other member services are ignored.
These services are important in maintaining cooperative membership and value, as
membership services are an incentive for producers to continue their patronage. Often,
proprietary ﬁrms offer a pay price that is greater than the cooperative blend price, to
compensate producers for the services these ﬁrms do not provide. Thus, dairy farmers
choose the type of compensation they will receive for their supply of raw milk. Farmers
may face a tradeoff, either in services from the cooperative, or in a higher blend price
ﬁ‘om the proprietary ﬁrm.

Cooperatives and proprietary ﬁrms operate at different stages of the supply chain
for ﬂuid and manufactured milk products. Bargaining cooperatives buy milk from farmer
members to sell raw milk to processors. Balancing cooperatives, in addition to having
the market options of bargaining cooperatives, possess the ability to manufacture Class
111 products (e.g., cheese, butter, nonfat dry milk powder) from raw milk supplies. In
contrast to cooperatives, proprietary ﬁrms commonly ﬁmction at more than one level of
the supply chain. Proprietary ﬁrms purchase milk directly ﬁom independent farmers or
cooperatives, process the raw milk into mostly Class I products, and distribute those

products to retail outlets. The processing and retailing activities of the proprietary ﬁrm

26

are ignored here
proprietary ﬁrm;
representative of
members. This r
a premium on C l
3.2.1 Model of
Proprieta
miners proPortic
representative pr
ﬂuid milk produ
farmers and dair
11 and Class 111
90 Percent of p!
Product OfIhe r
“Perez:

ﬁrm purchases

are ignored here as this research concentrates only on the selling activities of raw milk by
proprietary ﬁrms. In this context, it is easier to view the proprietary ﬁrm as a
representative of independent farmers, much like a cooperative represents cooperative
members. This representation is realistic as all PEC organizations operate to pay farmers
a premium on Class I and Class II milk.

3.2.1 Model of Proprietary Firms

Proprietary ﬁrms are businesses that return net income or proﬁts to investors and
owners proportional to their investment or share of ownership. In this research, the
representative proprietary ﬁrm is modeled as only procuring and distributing Class I, or
ﬂuid milk products, made from raw milk supplied by independent Michigan dairy
farmers and dairy cooperatives. Although Michigan proprietary ﬁrms manufacture Class
II and Class 111 products as shown in Table 3.2, Class I products usually account for over
90 percent of proprietary ﬁrm total sales. Thus, the ﬁrm’s total revenue is largely the
product of the retail price of ﬂuid milk and the quantity sold to the market.

Whereas cooperatives guarantee a market for all members’ milk, a proprietary
ﬁrm purchases only those quantities needed. The ability to choose quantities of milk
purchased and sold, based on market conditions gives proprietary ﬁrms an advantage
over cooperatives. During winter months when production reaches its lowest of the year,
and ﬂuid consumption peaks, proprietary ﬁrms turn to cooperatives to supply the raw
milk needed to meet the difference between market demand and quantities supplied by
independent producers. By creating this balancing relationship, proprietary ﬁrms
minimize balancing responsibilities and costs, and transfer these market responsibilities

to cooperatives.

27

To maxii
purchased. Qrp -
quantity of ﬂuid
proprietary ﬁrm

monthly by the 1L

 

proprietary ﬁrms

level. All produc

 

 

 

market at prices 1

m’th respect to thc
(3.2) 1
The prOpri-mm-y f

Quantity of O n

for 13.2) is:
(3.3") é

This result is th

To maximize proﬁts, proprietary ﬁrms ﬁ'eely choose the quantity of Class I milk
purchased, Q,” , as an input ﬁ'om independent farmers or cooperatives to equal the
quantity of ﬂuid milk delivered to the market, Q, . The price of Class I milk to the
proprietary ﬁrm, P,” , is greater than or equal to the minimum Class I price deﬁned
monthly by the federal orders, because P,” includes over-order premiums. The price
proprietary ﬁrms receive for ﬂuid milk, P, , is deﬁned by market conditions at the retail
level. All productive inputs, X , , excluding raw milk, are purchased in a competitive
market at prices 1,. Using these assumptions, the proprietary ﬁrm maximizes proﬁts

with respect to the quantity of milk purchased such that:
(32) Héapxl—Ip:Prgp-_ZI:Xr—PIPQIP

The proprietary ﬁrm chooses the quantity of Class I milk purchased, Q,” , to match the

quantity of Q, needed to supply the retail market at price P,”. The ﬁrst order condition

for (3.2) is:

(3.3) 6%Qf =1), —P,” =0.

This result is the expected microeconornic outcome for a proprietary ﬁrm that maximizes
proﬁts by choosing the quantity of Class I milk such that marginal revenue, P, , equals
marginal cost, P,” . The proprietary ﬁrm responds to minimum and market price changes
by adjusting the quantity of milk purchased and sold.

3.2.2 Models of Milk Cooperatives

To model bargaining and balancing cooperatives this research builds off of the

work of Helmberger and Hoos. Their work adapts the theory of the ﬁrm to the structure,
28

organization. ant
proﬁts to zero (j i.-
cooperatives ma‘
supplied by farm

Michigan

. i
and the quantity t

 

farms. Since qua

 

cooperative price
calculating the co-
Because 11
cooperatives to p:
However. Since a

organizations f0 r

organization, and principles of a cooperative. Assuming that a cooperative equates
proﬁts to zero (i.e., return all proﬁts to members through the blend price) they show that
cooperatives maximize the cooperative surplus through the price, given the quantity
supplied by farmers is ﬁxed.

Michigan dairy cooperative’s market actions are constrained by their membership
and the quantity of milk they control. Cooperatives accept all milk produced by member
farms. Since quantity is given through the independent choice of members, the price
cooperatives negotiate for raw milk is then the choice variable for a cooperative. The
cooperative price received for Class 1, Class II, and Class III milk is then used in
calculating the cooperative blend price.

Because federal orders view cooperatives as a single producer, they allow
cooperatives to pay members at class prices less than the federal order minimums.
However, since a cooperative competes with other cooperatives, as well as proprietary
organizations for milk supplies, competition and market conditions limit the potential to
pay less than federal order minimums prices.
3.2.2a The Bargaining Cooperative

Bargaining c00peratives direct raw milk to proprietary ﬁrms while negotiating a
price usually higher than the federal order minimums on behalf of its members. By
deﬁnition, bargaining cooperatives do not operate manufacturing or processing facilities
and are only involved in price negotiations.

A bargaining cooperative controls the disposition of raw milk supplied and directs
its supply to various outlets to achieve the highest net price. Bargaining cooperatives are

usually unable to sell 100 percent of milk supplies to Class I buyers. In the data analyzed

29

here. the represt
total milk for CI
manufacturing p
assumes that the
agency-in-comm

order minimum 1:

 

 

In this mo
Proprietary firms
C00perative of P
maximizes the qt
bargaining COOpe
and the national l

The cost 1
total quantity su,

blend price p; i

“here Q5 k
L ,9
re

here, the representative bargaining cooperative typically sold approximately 60 percent of
total milk for Class I uses, with remaining milk sold for Class H and Class III
manufacturing purposes. Due to the collective action between cooperatives, this research
assumes that the bargaining cooperative with or without an operational marketing
agency-in—common, attains a higher price for Class I and Class II milk than the federal
order minimum prices.

In this model, the bargaining cooperative sells the majority of its milk supply to

proprietary ﬁrms for use in the Class I market, Q ,” , at a price to the bargaining
cooperative of P,”. The bargaining cooperative sells all milk remaining, after it
maximizes the quantity sold for Class I use, to Class II and Class III buyers. The
bargaining cooperative receives a price P: for the quantity of milk sold for Class 11, Q,”, ,
and the national price P,,, for milk sold for use in manufactured dairy products, Q,”,,.

The cost to the bargaining cooperative of the milk supply is the product of the
total quantity supplied by cooperative members, Q',’ , and the bargaining cooperative’s
blend price Pg, illustrated in (3.1). The blend price in turn, is a function of the
cooperative class prices and quantities. Other productive inputs needed, X, , are
purchased in a competitive market at prices 1,. The bargaining cooperative’s proﬁt
equation is:

(3.4) 1]” = B”Q7(P,”) + 12191: (on + 1’”,er —ZI.X.~ mm: 1 mono:
where Q; = Qf + Q}; + Q7”. P,” and P5 include any over-order premiums collected by

the cooperative alone or with other milk marketing organizations. To account for

competition ﬁom nearby surplus milk markets, the bargaining cooperative’s negotiated

3O

Class I price is c

where P,’ is the

transportation frt
bargaining coopc
from a surplus m
communication v

constraint also ex

 

 

The total (
farmer decisions
cooperatives can
this r Esearch ign.

Profits to memb

Class I price is constrained by:
(3.5) P” s P," +13,
where P,’ is the cost of Class I milk in the nearest surplus market, and P, is the cost of

transportation from the surplus market. This constraint on price is added as the

bargaining cooperative cannot price Class I milk above the cost that milk can be imported

ﬁom a surplus market. However, the bargaining cooperative can affect P,’ through

communication with cooperatives in the surplus market, who set Pf. The same
constraint also exists for Class II milk.

The total quantity of milk, Q; , is exogenous and is determined by individual
farmer decisions rather than being controlled directly by the cooperative. Of course,
cooperatives can direct long-run production through pricing and beneﬁt decisions, but
this research ignores these possibilities. Since cooperatives operate at cost and return
proﬁts to members through the blend price, the proﬁt equation is set equal to zero and

rearranged to yield the bargaining cooperative’s optimization problem:

(3.6) pagPtoﬁ = new» aroma?) + mo?” -ZI.X,— ,

I‘ll

subjectto P,”SP,‘+P andl’,§’51’,j+P,.

To maximize the bargaining cooperative’s blend price, given that P,” and P; are greater
than P,,, , bargaining cooperatives maximize revenue generated through the sale of milk
into Class I and Class 11 markets. The quantity of milk directed to each market is not a
choice variable of a cooperative. Rather, once a cooperative announces its Class I and
Class II prices, proprietary ﬁrms react by purchasing the corresponding quantity of milk
for each class from their individual demand curve.

31

The ﬁrst

11. is:

(3.7)

[)5
. , b = “1'
“er6 E, (f

5.:

 

condition indicatc
price. sets margir
Where the elastic
pricing actions 0
from a surplus n
581 its Class I or
ﬁrm could 1mm

3.2.2b Ba la n ci

The ﬁrst order condition for (3.6) for milk sold to Class i, for i = Class I or Class

 

II, is:
6P3Qﬁ _ 6Q.” _
(3.7) ggrgfuﬂ 01'3” -0
P.” .
:QIb+Qf[QIbI@’%b)=O
=Qf’(l+E,”)-—0,
P." aQb
where E,” = 5; ( %P”] isthe price elasticity of demand for Classimilk. This

condition indicates that a bargaining cooperative, to maximize the cooperative blend
price, sets marginal revenue equal to zero for Class I and Class II milk which occurs
where the elasticity is negative one. However, a bargaining cooperative is constrained by
pricing actions of its competitors, as well as the cost of transporting milk into the market
from a surplus milk market. These constraints imply that a bargaining cooperative cannot
set its Class I or Class 11 price higher than its competitors, or the cost that a proprietary
ﬁrm could import milk ﬁom a surplus market.
3.2.2b Balancing Cooperative

Balancing cooperatives perform the ﬁrnctions of bargaining cooperatives but also
possess manufacturing capacity for Class 111 products. In addition to deciding
cooperative class prices, a balancing cooperative can potentially choose the quantity of
milk sold to proprietary ﬁrms because it operates manufacturing facilities to convert raw
milk into storable manufactured dairy products. Balancing cooperatives typically market

these manufactured dairy products directly to wholesale markets. The operations of

32

balancing COOP‘
cooperative C lib
The mOci
bargaining coop
from manufactur
quantity of stora
price captures lht

above P,,, from 5

 

chain.

Balancing
the COOperatiVe b
raw milk (“Tenet
Class

I“ buyers.

to the productic

OI‘S. PYOYXXS ﬁ

Ofa “mincem

balancing cooperatives are constrained by market balancing, marketing costs, and
cooperative class prices.

The model for the balancing cooperative is identical to the model for the
bargaining cooperative with the exception of an added choice and subsequent revenue
from manufactured dairy products. This additional source is represented as selling a

quantity of storable manufactured products, S, in a national market at price P,. This

price captures the additional revenue, positive or negative, generated by the cooperative
above P,,, from selling these products to market and expanding its role in the supply
chain.

Balancing cooperatives operate at cost and return proﬁts to members to maximize

the cooperative blend price. The balancing cooperative therefore chooses the quantity of

raw milk directed to the ﬂuid market, Q,” , and the quantity sold to external Class II and
Class III buyers, Q}; and Q,’;‘, , as well as the quantity used internally in S storable
products, given the quantity of raw milk supplied by members, Q," . Milk that is directed

to the production of S is accounted for in Q,”',,. This quantity of milk is reported as being

sold internally, as the blend price is based on Class III utilization in making the products
of S. Proﬁts from the sale of S are returned to cooperative members annually in the form
of a “thirteenth” check. The thirteenth check represents proﬁts the balancing cooperative
received ﬁ'om supplying manufactured dairy products to the national market. The
thirteenth check terminology is derived ﬁom a payment to cooperative members above
and beyond the twelve regular monthly milk checks.

A balancing cooperative maximizes proﬁt over three choice variables, the price of

milk sold for Class I use, the price of milk sold for Class 11 use, and the quantity of milk

33

balanced or ”58
COOperative. Om
are based on ind
cooperatives em
for butter. 130W

specific month is

 

cooperative can
cooperative rarel
obligations. beca
Class [1 uses. I‘lt
manufactured prr
than demand. Tl
to meet its suppl
l5 SUPPlIEd by its
facilities and c0]
Using th
indicator to m f,

cooperative is:
(3 .8)

“here a
Q; §9

21

balanced or used in the production of S. Again, as modeled with the bargaining
cooperative, once a cooperative sets its class prices, quantities sold to a particular class
are based on individual demand curves of proprietary ﬁrm buyers. Balancing
cooperatives enter into long-term contracts with buyers of manufactured dairy products
for butter, powder, and cheese. Thus, the quantity of milk needed to produce S in a
speciﬁc month is usually known by the balancing cooperative months in advance, so the
cooperative can meet its contractual supply obligations. However, the balancing
cooperative rarely chooses to balance more milk than what is necessary to meet its supply
obligations, because it desires to maximize the quantities of milk sold for Class I and
Class H uses. However, a balancing cooperative may be forced to convert more milk into
manufactured products than what is contractually needed if market supplies are greater
than demand. This later case occurs after a balancing cooperative converts enough milk
to meet its supply obligations, and sells less milk to Class I and Class II buyers than what
is supplied by its members. This residual milk is then also directed to the manufacturing
facilities and converted to storable dairy products.

Using the same notation as the bargaining cooperative, changing the ﬁrm
indicator to m for the balancing cooperative, the proﬁt equation for the balancing

cooperative is:

(3 8) II .. = P 1.9,. (P 1"“ ”17' 917 (P17 >+ Prion + as
- 2 17X.- - C(S) — Farmers 1 P,,,,Qrior,
where Q," = Q,” + Q}; + Q3, , and C is the total cost of production for S. Balancing

cooperatives also account for competition from surplus markets for Class I and Class II

34

milk and again

Given that the b

members throng

m:

(3.9) 5’"

Sl

 

 

 

As with the barg
blend price. The

or Class II. are:

(3.10) 51

(3.11) a

{3.10) is iDIErprc

prOdUCI io n The

COUStr

milk, and again assume that the price choice is constrained by potential imports, P,’ , that
Prsnwa and PJ's—aha.

Given that the balancing cooperative sets proﬁt equal to zero, by returning proﬁts to their

members through the blend price, the balancing cooperative’s optimization problem is:

312%” P37Qi" = PImQIm(P1m)+ P17Q17'(P17)+ PIIIQIMII + 1’35
(3.9) " ’”
—Z IiX1'—C(S)9

subject to P," SP,’ +P, and 1),," _<.P,j +P,.
As with the bargaining cooperative, the balancing cooperative maximizes the cooperative
blend price. The ﬁrst order conditions for (3.9) for milk sold to Class i, where i = Class I

or Class II, are:

6P’"Q"'/_ _aC -
(3.10) B! TBS—P, /ag_0, and

(3.11) ”P879%P,, =Q,’"(1+E,”')=O for Classimilk.

(3.10) is interpreted as marginal revenue from S, P, , equals the marginal cost of

production. The result in (3.11) is identical to the bargaining cooperative, where the
balancing cooperative maximizes total revenue by setting marginal revenue equal to zero
which occurs at the point where the elasticity of demand is negative one. The pricing
constraints on the balancing cooperative also remain the same, speciﬁcally that its Class I
or Class II price cannot be higher than its competitors or the cost of importing milk from
a surplus market.

This chapter has set the stage to analyze behavior and interaction of Michigan

milk marketing organizations in the context of the PEC. The ﬁrst section accounted for

35

the institutions
considered in If
provide an undc
research next re
1998. focusing t

marketing agenc

 

the institutions and policies that affect ﬁrm conduct, and then these factors were
considered in the construction of individual ﬁrm optimization problems. The results
provide an understanding of how ﬁrms should behave in a competitive milk market. This
research next reviews actual behavior of Michigan ﬁrms for the period of 1980 through
1998, focusing on joint proﬁt maximization through the operation of a ﬂuid-milk

marketing agency-in-common.

36

 

Chap

Underst;
ofthe structure .

a historical acco

 

common in 195.1
market. in temis

4.1 Establisl

 

To increa
Where total rever

formed the nati01
over-order premi
Participating Coc
Which included :
ordﬂ minimUm
agency'in‘COmn
In the la'
MusltEggn and ,
DelTOli Order 3)
created to COVE

NIiChiga“ 0rde

Chapter 4: Michigan Fluid-Milk Marketing Agencies-In-Common
and Market Performance

Understanding the over-order premiums generated in Michigan requires a review
of the structure and performance of the Michigan milk market. This chapter begins with
a historical account of the events leading to Michigan’s ﬁrst milk marketing agency-in-
common in 1956. The effects of a marketing agency-in-common on Michigan’s milk
market, in terms of prices and quantities supplied, are examined in the second section.
4.1 Establishment and Growth of the Producers’ Equalization Committee

To increase farm-level milk prices in the Class I market with an inelastic demand,
where total revenue increases with an increase in price, Michigan dairy cooperatives
formed the nation’s ﬁrst marketing agency-in-common or over-order pool to generate
over-order premiums in April 1956 in the Detroit market. In this initial month,
participating cooperatives negotiated a Class I price of ﬁve dollars per hundredweight
which included a 52 cent over-order premium, deﬁned as the amount above the federal
order minimum Class 1 price, which was pooled and shared between all marketing
agency-in—common members.

In the late 1950s, Michigan cooperatives established over-order pools in the
Muskegon and Upstate Orders using the Detroit pool as a model. In 1960, when the
Detroit order expanded across southern Michigan, four new over-order pools were
created to cover the separate metropolitan markets in this area. In 1963, these ﬁve
separate over-order pools were merged to create one “superpool” for the entire Southern

Michigan Order.

37

Historic
proprietary mi ll
the Michigan so
determined thes
operating in the
Committee whic
the control of on

the Southern M It.

 

In 1966. t
the Producer‘s E1
regional marketii
Originally starter
marketing coope
sew to include 1

From 19-
Order Premiums
CmperallV'es be
”turned to farm

ihe GLAIF I0 &

Figufe .

Historically, Michigan cooperatives, with little input or negotiation from
proprietary milk processors, announced monthly ﬂuid-milk prices. Prior to 1963, when
the Michigan superpool was formed, local bargaining committees in each area
determined these over-order prices. In 1961, representatives of three major cooperatives
operating in the Southern Michigan Order formed the Michigan Cooperative Bargaining
Committee which brought over-order pricing in the ﬁve regional over-order pools under
the control of one body. The Michigan Cooperative Bargaining Committee set prices for
the Southem Michigan over-order pools and superpool until 1966.

In 1966, the Michigan Cooperative Bargaining Committee changed its name to
the Producer’s Equalization Committee. At about the same time, PEC members joined a
regional marketing agency-in-common, the Great Lakes Marketing Federation (GLMF).
Originally started as an association to share information, the GLMF consisted of dairy
marketing cooperatives in Michigan, Ohio, Pennsylvania, and Indiana. Later the GLMF
grew to include cooperatives in 15 contiguous states, ﬁom Michigan to Florida.

From 1966 to 1984, the GLMF assumed the role of setting class prices and over-
order premiums for the markets in which member cooperatives participated.
Cooperatives believed that a regional bargaining association effectively increased prices
returned to farmers and coordinated ﬂuid-milk sales between markets. The PEC relied on
the GLMF to set the class prices and over-order premiums for the Michigan market.

Figure 4.] displays the major events in the development of Michigan marketing
agencies-in-common since 1956. The tirneline also indicates the three PEC price wars

that are reviewed, in detail, in Chapter 5.

38

 

 

Follovvin
COOperatives ina
the PEC resume
and announces z
PEC members L
Purposes_ Since
PEC members c

In 1992,
volumari]y 30in
meaning Premir
the C1355 II mil]
ﬁrm paniclpalic
Imilk “Ere add.

I00 DErcem 0f 1,

Piﬁlcipat io n 0

 

Figure 4.1 Timeline of the Michigan Milk Cartels

 

1997 Price
Area Pools Merge WE}?
1963 1982 Price 1992 Price i
Detroit Pool Forms : PEC Created War War I
Aprill956 : 1966 , : ,
' ' ' I I i
I i 1 i i J 1 : 1 L :
19'50 1960 1970 1980 1990 20'00
Area Pools Proprietary
Operate GLMF Sets Class Prices Firms in PEC
H O 9 .———O

 

 

 

 

Following the termination of the GLMF in 1984—caused by Southern
cooperatives inability to control sufficient milk supplies to collect the GLMF premium—
the PEC resumed Michigan over-order pricing on its own. Today, the PEC determines
and announces a monthly price and over-order premium for Class I and Class II milk.
PEC members use these prices to sell milk in the Michigan market for Class I or Class II
purposes. Since milk processors in the long-run switch milk suppliers based on price,
PEC members cannot effectively set an individual ﬁrm price higher than the PEC price.

In 1992, the milk controlled by proprietary milk processing ﬁrms effectively and
voluntarily joined the PEC. Prior to 1992, only cooperatives participated in the PEC,
meaning premiums from approximately 71 percent of the Class I milk and 98 percent of
the Class II milk were divided over 84 percent of total market milk. With proprietary
ﬁrm participation, premiums generated from the sale of the remaining 29 percent of Class
I milk were added to the marketing agency-in-common, and subsequently divided over
100 percent of total market milk production. To illustrate the impact of proprietary ﬁrm

participation on the PEC, consider a Class I premium of $1 .00 per hundredweight. For a

39

 

  
   
   

proprietary ﬁm
Later in
between June 1‘
prior to 1985. n.
1980. In this ne
price was. A p
agreement alter
premiums. In t
Where price eqi
Despite these c
Periods, or l7§
PTEmiUms_
I“figure
temB‘Dﬁcer
1998. in tea,

DECEmbel. 1 9,

typical month in 1998 with 170 million pounds of Class I milk sales, participation by
proprietary ﬁrms increases monthly PEC revenues by approximately $493,000.

Later in this research, empirical work concentrates on the operation of the PEC
between June 1985 and December 1998. However, since the history of the PEC is known
prior to 1985, narrative portions of this research describe its activity beginning in May
1980. In this nearly 19-year period, the PEC’s operation was interrupted three times by
price wars. A price war occurs when a participating PEC member deviates from the
agreement altering either the collection, pooling, or distribution of market over-order
premiums. In two of the price wars, PEC members reverted to the competitive outcome,
where price equals the minimum federal order class price plus a small service charge.
Despite these disruptions, the PEC operated for 80 percent of the potential pricing
periods, or 179 of 224 months, and successfully collected and pooled Class I and Class II
premiums.

Figure 4.2 displays the PEC Class I premium on a per hundredweight basis in real
terms—December 1998 dollars—and nominal terms, from May 1980 through December
1998. In real terms, Class I premiums range from a high of $2.89 per hundredweight in
December 1990 to a low of zero during each of the three price wars when the PEC
agreement was void. In nominal terms, Class I premiums range from a high of $1 .87 per
hundredweight in December 1990 to a low of zero the three price wars. For a period of
almost three years, the PEC collected a constant premium of $1 .00 per hundredweight.

However, during this time the premium steadily decreased in real terms.

40

. ‘ t
Premium (Mew )

 

1981‘

T0 furtl
weighted aver;
using the Iota}
the minimUm ,
Pounds OfCIa:

using the “Big

6x1"63M in r

Figure 4.2 PEC Class I Premiums from May 1980 through December 1998

 

 

 

 

 

3.00 4
2.50 . . _
0' i \
A I:. I, :‘I
E ”-00 “ ' 5‘; iii
a '2 3' J, ;:t
V " I. h .
1.50 . I: z: 3w“: , ; ,: .
g 3:, ",- { {PL-4J1.
on 0 .‘I‘ o
E 1.“) ,i\,/{",:¥’o',:rﬁ, ’5 ' H: o. I |~ ..‘\\
0.50 4
0'm I I I I I T I I I I I I I I V I I I
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998
Year
------- RealPremiumDec.98 NominalPremimn

 

 

 

 

 

 

 

To further illustrate the magnitude of the PEC premiums, Table 4.1 reports the
weighted average annual premium for Class I milk. The weighted average was calculated
using the total quantity of milk sold by the PEC in each use class. The table also shows
the minimum and maximum premium collected in each year. This table also reports the
pounds of Class I milk marketed through the PEC. Total Class I revenues are calculated
using the weighted average premium and the total Class I milk quantities. All values

expressed in Table 4.1 are expressed in constant December 1998 dollars.

4]

Table 4.1 PEC:
Year

—-———-—-'—'1‘

1980
1981
1982 '
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998

\—

 

 

 

 

The PEC
IIICIEaSeS markel

[hrOUgh Class II

Figum 4.

Table 4.1 PEC Class I Premiums and Descriptive Statistics

 

 

 

 

Year Avg. Prem. Min. Max. Class I Sales Class I
(WM) (WM) (S/CWI) (lbs) Revenue

1980 1.18 0.92 1.38 - -

1981 - 0.89 1.13 - -

1982 - 0.00 1.05 - -

1983 0.00 0.00 0.00 - -

1984 0.00 0.00 0.00 - -

1985 - 0.00 1.01 - -

1986 0.95 0.84 1.06 1,818,103,374 $17,271,982
1987 0.95 0.83 1.10 1,732,492,294 $16,458,677
1988 1.39 0.89 2.05 1,695,522,807 $23,567,767
1989 1.07 0.72 1.31 1,656,491,922 $17,724,464
1990 1.42 0.67 2.89 1,625,975,079 $23,088,846
1991 1.80 1.47 2.38 1,630,325,870 $29,345,866
1992 - 0.00 1.68 1,818,103,374 -

1993 1 .41 1.33 1 .48 2,014,397,109 $28,402,999
1994 1.40 1.12 1.81 2,007,188,223 $28,100,635
1995 1.18 1.02 1.45 2,065,239,978 $24,369,832
1996 1.07 0.97 1.35 2,070,371,893 $22,152,979
1997 1.21 0.00 2.51 1,912,755,193 $23,144,338
1998 0.74 0.00 1.52 1,939,290,417 $14,350,749

 

4.2 The PEC and the Michigan Milk Market

The PEC operates in Michigan to increase the price of raw milk. The PEC
increases market revenues through increased Class I pay prices and to a lesser extent
through Class 11 pay prices.

Figure 4.3 graphically depicts the performance of this market including minimum
prices, PEC premiums, and quantities of milk sold, using as an example the prices and
quantities for May 1997 with prices expressed in December 1998 terms. The ﬁgure

contains three demand curves, D, , D,, , and D,,, , representing the demand for Class 1,

Class II, and Class HI milk. The demand elasticities vary, with ﬂuid demand being very

inelastic, Class II milk more elastic, and Class III demand perfectly elastic. As discussed

42

in Chapter 3. C

thus is modeIe

  
 
 

The Mu
are $19.82 and

organizations 11

the quantity of r

 

Figure 4.3 The
23.00
22.00

P

21.00 T
, ..

A -000 f

E

3 19.00

‘1‘; r

a _

.2 18.00

a I

a. -
17.00 .
I600
15.00 -
l4.00 -

in Chapter 3, Class III milk used in manufactured products faces a national market and

thus is modeled as perfectly elastic demand.
The May 1997 minimum federal order Class I and Class 11 prices, P,“n and P,;“’"

are $19.82 and $17.80. In the absence of the PEC, competition between milk marketing
organizations forces class prices to the federal order minimums as there is always excess
milk available for the higher-priced uses. The corresponding competitive equilibrium for
the quantity of milk demanded at the minimum class prices are indicated as Q,‘ and Q,°, .

Figure 4.3 The PEC Effect on Market Performance for May 1997

 

 

 

 

 

23.00 —,
22.00 - D'
It? ..........................
21.00 -‘ DH : \\
l \
\
20.00 ~ pm ! \
g --------------------------- 1--- 1.
: ' -. \\ MC=S
v H --------------------------- l— ----- b ~~~~~~~~~~
.3 18.00 ‘I'an-n i I . \ -\‘.\ ~' ~
6: L """""""""""""" 7“"? """ _ “’. "--~...‘_’9.u:, 1
. I I -----
17.00 « i , I : t,‘
: . . 1
1600 l I ’ i ‘ '
° 4 i . f : I I
: i i l i I
15.00 1 Dr" ! I 1 1 = '
6 I < l
‘I c' d. c: | Q
14-00 "““”‘ r —T' T ﬂ 1 101 J 1 Ql—I‘wg‘h‘rQ‘g.‘ 1 T 1 1 2' QT
c
25 75 125 175 225 275 325 375 QT 425
Milion Pounds of Milt

 

 

 

The PEC collects an over-order premium and raises the price of milk sold for

Class I and Class 11 uses. The Class I and Class II PEC prices in Figure 4.3 are labeled

Epic and P5“. The PEC Class I and Class 11 prices in this month were $21.21 and

43

  
 

$18.41 reprCSC'

c1333 1' milk.

The qua
Q 15 at the in
these prices. ab

were demandec
and the quantit
utilizationxsr

Total r

intersection 01
fannefS. PEC

prOClUCIIOn of

p rices. Total

milk.

The a]

SPEC IﬁCal I)“ ,
However, the

genewted b\ I

the
COmpet II I \_

H milk is su fti

$18.41 representing a $1.39 and $0.61 over-order premium. The over-order premium for

Class 1' milk, p, , for i = Class I or Class II, is deﬁned as the difference between the PEC
price and the federal order minimum price, or RPEC — Pf“ = p,.
The quantity of milk demanded by class when the PEC operates, labeled Q; and

Q], , is at the intersection of the demand curves and the corresponding PEC prices. At

these prices, about 170 and 73 million pounds of Class I and Class II milk respectively,
were demanded. The difference in the quantity of milk demanded at the minimum prices
and the quantity demanded at the PEC prices—speciﬁcally less Class I and Class II
utilization—shifts to the Class 111 national market.

Total milk produced in periods of competition and PEC operation is found at the

intersection of the blend price and the marginal cost, or supply curve, for Michigan dairy
farmers. PEC rents increase the farm-blend price from P5“ to Pg“ , and lead to the

production of more milk than what would be supplied at the federal order minimum
prices. Total market demand during operation of the PEC was 394 million pounds of
milk.

The arrows in Figure 4.3 indicate the PEC effect on the Michigan milk market.
Speciﬁcally, Class I and Class 11 sales decrease, and total milk production increases.
However, the PEC successfully increases total market revenue. The total revenue
generated by the product of the PEC blend price and output is greater than the product of

the competitive blend price and competitive output because demand for Class I and Class

II milk is sufﬁciently inelastic. Speciﬁcally, (Pgh, XQ;)> (P5,, XQ; ), as the increase in

44

total market rex

   
  

utilization undc
The hen

milk marketing

maximize mark<
common are cle
Opted to not cot
order premium

surrounding 1h

total market revenue more than offsets the increasing weight of a higher Class III
utilization under the PECs’ operation.

The beneﬁts to milk marketing organizations of forming and operating a ﬂuid-
milk marketing agency-in-common are obvious. Michigan milk marketing organizations
were some of the ﬁrst to translate these opportunities into regional over-order pools to
maximize market revenues. However, although the gains from marketing agencies-in-
common are clear to all industry participants, in recent history ﬁrms have periodically
opted to not cooperate and instigate a price war rather than collect and pool these over-
order premiums. The next chapter reviews the recent history of the PEC and events

surrounding the three recent PEC price wars.

45

   
 

Using.“
the causes and
examines thOSC
PEC reformed
agreement and

Organiz

The representa‘
bargaining C00
manufacturing
consequences
organizations.
5.] PEC 1
Price ‘
existing PEC
[I premiums.
cooperate. or
PUFpoSes, a p
001 jointly p0

dElinition 3134

I

mbers con

action by On
. e

O .

Chapter 5: Recent History of the PEC: The Price Wars

Using accounts of participants reported in trade publications, this chapter reviews
the causes and effects of each of the three PEC price wars since 1980. The ﬁrst section
examines those factors leading to each price war and the resulting agreement once the
PEC re-formed. The second section reviews the pooling mechanism of each PEC
agreement and chronicles its changes due to the price wars.

Organization labels are employed in this research to denote PEC participants.
The representative aggregate proprietary ﬁrm is denoted Firm P, the representative
bargaining cooperative, Cooperative B, and the representative balancing or
manufacturing cooperative, Cooperative M. This chapter attributes actions and
consequences of the price wars and PEC agreements to these aggregate milk marketing
organizations.
5.1 PEC Price Wars

Price wars are deﬁned as any time at least one PEC member deviates from the
existing PEC agreement. Upon such an action, the PEC stops pooling Class I and Class
II premiums. This deﬁnition does not necessarily mean that PEC members do not tacitly
cooperate, or collect over-order premiums individually during a price war. For our
purposes, a price war merely means that the PEC, as it existed prior to that period, does
not jointly pool and transfer premiums to all of the previous PEC members. This
deﬁnition also encompasses the exit of one member, even if the two remaining PEC
members continue to collect and pool premiums. Therefore, a price war refers to any
action by one PEC member that alters either the collection, pooling, or transfer of market

over-order premiums.

46

PEC mt

 

distribution of

 

well as delineat
Table 5
Premium. mini
Cause, and resr
market loss is
WT hundredwc

quantities are

40.

PEC members initiate a price war in an effort to increase short-run revenues, alter
distribution of collected rents, change membership, or improve negotiation position.
Since 1980, three price wars have prevented market-wide pooling of Class I and Class II
premiums for a total of 46 months. These price wars resulted in a structural change to the
PEC agreement. Each PEC agreement deﬁnes the pooling mechanism which includes
provisions for collecting, pooling, and transferring total PEC rents between members, as
well as delineating governance of the marketing agency-in-common.

Table 5.] summarizes each price war in terms of the initial drop of the Class I
premium, minimum Class I premium, duration, total market revenue foregone, initial
cause, and results. Reported premiums are on a per hundredweight basis. Estimated
market loss is calculated using a ﬁxed Class I and Class 11 premium of $1 .00 and $0.40
per hundredweight, inﬂated to constant December 1998 dollars. Class I and Class II
quantities are estimated using monthly average utilization and total production in Order

40.

47

Table 5.1 Characteristics of PEC Price Wars

 

 

PEC Price War 1982 1992 1997
Initial Premium Drop $0.61 $1.18 $0.22
Minimum Premium $0.00 $0.00 $0.78
Duration (mo.) 32 6 8
Market Revenue Foregone - $14,347,555 $3,497,831
Initial Cause Co-op B desires Co-op B desires Firm P desires to
change in change in distribution retain collected Class
distribution of of PEC rents. I and Class II
PEC rents. premiums.
PEC Results No change to Firm P joins PEC; Members retain
pooling Co—ops pool cheese higher portion of own
mechanism; milk premiums; Class I premiums;
Voting structure Members retain MBF eliminated.
changes. portion of own Class
I premiums; MBF
created to

compensate co-ops
for market balancing.

Traditionally, Firm P and Cooperative B transfer more to the PEC in Class I and
II premiums than each organization’s respective settlement, or net payoff from the PEC.
This occurs because they have a high Class I utilization and do not balance the market.
Conversely, Cooperative M usually receives a larger net payoff from the PEC than what
it individually collects in Class I and Class II premiums. As the next section reveals, this
net pay situation has been the cause of strife between PEC members contributing to the
volatile nature of this agreement. The three price wars between 1980 and 1998 illustrate
this instability.

In 1982, Cooperative B withdrew from the PEC in November (Michigan Milk
Messenger). The 1982 price war lasted almost three years until the PEC re-formed in

July 1985.

48

In June 1992, Cooperative B again terminated the PEC agreement for six months
(Michigan Farmer). The PEC resumed operation in December 1992 after members
agreed to alter the pooling mechanism. Perhaps the most important change following the
1992 price war, Firm P voluntary agreed to join the PEC and pool Class I and Class II
premiums with the two cooperatives.

Most recently, the PEC collapsed in September 1997 with Firm P’s exit
(Michigan Farm News). Although both cooperative PEC members continued to pool
premiums for eight months before Firm P voluntarily returned, this period constituted a
price war as premiums were not pooled market-wide. To understand the stability and
incentives of the PEC, each price war is described in detail below.

5.1.1 The Price War of 1982

In November 1982, Cooperative B announced they would not pool over-order
premiums because efforts to change the PEC pooling mechanism were not supported by
Cooperative M. The PEC did not operate from November 1982 to July 1985. The 1982
price war resulted in two different states of market interaction between former PEC
members.

Initially, PEC members continued to collect, but not pool over-order premiums at
the same level as the PEC on Class I and Class II milk in a state of tacit cooperation.
This cooperative state lasted from November 1982 to October 1984 and favored
Cooperative B as it had a higher Class I utilization than Cooperative M. Using a Class I
premium of 90 cents, ignoring Class H milk, on average Cooperative B collected about
$357,000 per month in Class I premiums during the cooperative state. When divided

over 85 million pounds of total milk, these premiums resulted in a 42 cent net payoff or

49

net return per hundredweight. When participating in the PEC, Cooperative B on average
received a $278,000 monthly net payoff which translated into a 33 cent net return per
hundredweight. These representative ﬁgures are based on the sale of approximately 40
million pounds of Class I milk for Cooperative B, representing a Class I utilization of 47
percent, with a 20 percent Class I market share.

On the other hand, Cooperative M’s net monthly payoff was less in the
cooperative state than it was under the PEC. Employing a Class I premium of 90 cents,
ignoring Class H milk, under the PEC Cooperative M typically sold 100 million pounds
of Class I milk and 300 million pounds of total milk resulting in a 33 cent net PEC payoff
per hundredweight, representing $981,000 in monthly revenues. In the cooperative state,
for the same quantities and premium level, Cooperative M collected $900,000 in
premiums which equaled a 30 cent net return per hundredweight. Cooperative M had a
33 percent Class I utilization and controlled 50 percent of the Class I market.

With unsuccessful negotiations between the cooperatives to reinstate the PEC,
member leaving Cooperative M to join Cooperative B with its higher pay prices, and
increasing competition for Class 1 milk sales, Cooperative M stopped collecting over-
order premiums in November 1984. As a result, the ﬂuid market essentially reverted to
competition, meaning over-order premiums were not collected on Class I or Class H
milk. To maintain competitiveness and retain its Class I buyers who were price sensitive,
Cooperative B followed Cooperative M’s lead as both organizations received only the
federal order minimum price plus a small premium for servicing the market. Cooperative

B generated approximately $52,000 per month in Class I premium revenues, 13 cents per

50

hundredweight, in this competitive state rather than the $357,000 from cooperative state
and $278,000 under the PEC (Michigan Farmer).

Michigan cooperatives agreed to a new PEC agreement implemented in June
1985 after the 32-month price war. Cooperative B returned to the PEC because net
payoffs from the PEC exceeded the net payoffs it could generate independently at the
competitive level. The 1985 agreement did not alter the pooling mechanism of the PEC,
as total premiums continued to be shared across all cooperative milk.

Following the demise of the GLMF in 1984, monthly PEC premiums were
determined using a price schedule until the 1992 price war (Table 5.2). PEC members
were assured a Class I premium between 58 and 72 cents per hundredweight. Even in the

new agreement, milk controlled by proprietary ﬁrms, Firm P, remained outside of the

 

 

PEC.
Table 5.2 PEC Pricing Schedule of 1985
If Federal Order 40 Class Then the Class I And the Class I
I Price ($/cwt) is between: PEC Price ($/cwt) is: Premium ($/cwt) is:
$12.30 and $12.44 $13.02 $0.58 - 0.72
$12.45 and $12.59 $13.17 $0.58 - 0.72
$12.60 and $12.74 $13.32 $0.58 - 0.72
$12.75 and $12.89 $13.47 $0.58 - 0.72
$12.90 and $13.04 $13.62 $0.58 - 0.72
$13.05 and $13.19 $13.77 $0.58 - 0.72

 

5.1.2 The Price War of 1992

In 1991, Michigan dairy farmers received the lowest milk price in more than 10
years. PEC members, believing cooperatives were losing income and membership to
proprietary ﬁrms, Firm P, they began an effort to force Firm P’s participation in the PEC

through state legislation. At that time, Firm P had an approximate 20 percent Class I

51

market share on about 30 million pounds of monthly Class I sales. Cooperatives argued
that these sales generated over-order premiums for Firm P, but Firm P did not pool these
premiums with c00peratives which had a higher Class II and Class III utilization. For
example, in January 1992, Firm P’s participation in the PEC would have increased PEC
revenues by $362,700 using the actual Class I premium of $1 . 1 7/cwt for the 31 million
pounds ofFirrn P Class I sales.

While drafting the language of a proposed state milk marketing order, debate
resumed between c00peratives regarding premiums that qualify for pooling and the best
manner to share collected PEC rents. Beginning in June 1992, after a year of
negotiations, Cooperative B withdrew from the PEC. Cooperative B cited the premiums
that were collected but not shared on Class III milk used to manufacture cheese as the
reason for their withdrawl.

Historically, premiums collected on the sale of cheese milk were not pooled, yet
the PEC included this milk in the calculation of the monthly PEC settlement. The
inclusion of cheese milk quantities, and exclusion of these collected premiums, lowered
the average monthly PEC blend premium paid on all milk. That is, if cheese milk was
excluded from the monthly calculations, Class I and Class II premiums would be divided
over a smaller quantity of milk increasing the PEC blend premium and payoffs to
Cooperative B.

Cooperative B argued for the inclusion of these Class III premiums, or exclusion
of Class 1H milk as Cooperative M collected these types of premiums. A high Class I and
Class II utilization by Cooperative B prevented it ﬁ'om entering into ﬁrll-supply contracts

with Class HI processors and regularly collecting Class III cheese premiums.

52

In December 1992, with momentum building for state legislation to force Firm
P’s PEC participation, the cooperatives resumed participation in the PEC after six months
without market over-order premiums. Firm P also voluntarily joined the PEC, an action
some attributed to the threat of the state marketing order that would have formalized
pooled premiums on Class I milk sales. The resulting PEC agreement included many
changes that addressed the issue of cheese milk premiums and enticed Firm P’s PEC
cooperation.

In creating the new 1992 PEC agreement, there was little dispute in arriving at the
desired level of Class I and Class II premiums (Michigan Farm News). However, the
milk qualifying for the PEC and the method of dividing total PEC premiums among
participants was a topic of much debate. Cooperative B, with a higher Class I utilization,
desired a minimum amount of Class I premiums pooled, and preferred all Class HI milk
be excluded from settlement calculations. This plan would have reduced Cooperative B’s
payments into the PEC, allowing the cooperative to retain a greater proportion of
collected premiums.

Conversely, Cooperative M, with a lower Class I utilization and greater Class 111
sales, wanted a maximum amount of collected Class I premiums pooled and all Class III
milk included in pool settlement calculations. This plan would have increased net
payoﬁs to Cooperative M from the PEC, and allowed the cooperative to retain premiums
they collected for Class III cheese milk.

To settle these differences and appease the demands of Firm P, the 1992 PEC
agreement departed from the previous pooling mechanism of dividing total premiums by

total milk. Instead, the 1992 agreement created three separate pools, Pool A, Pool B, and

53

a Market Balancing Fund (MBF). Essentially, Firm P received payment ﬁ'om Pool A,
cooperatives ﬁ'om Pool B, and the premiums from the MBF were distributed to
cooperatives for balancing the market. The purpose of the MBF was to eliminate free
rider aspects in the market and pay for the maintenance of balancing plants at a cost to
cooperative membership. The details and effects of the pooling mechanism created by
the 1992 PEC agreement are reviewed later in this chapter.
5.1.3 The Price War of 1997

After ﬁve years of voluntarily participation in the PEC, Firm P exited the
organization in September 1997. PEC cooperatives continued to collect and pool
premiums on Class I and Class II milk sales, while Firm P collected but did not pool the
same premiums, free-riding on participating PEC cooperatives’ bargaining. During the
price war, Firm P increased ﬁrm-level proﬁts and its pay price to Michigan dairy farmers
who sold milk directly to Firm P. However, in an aggregate sense, Michigan dairy
farmers belonging to participating cooperatives lost revenues from the premiums not
pooled by Firm P.

Following the initial withdrawal of Firm P, PEC cooperatives renewed efforts to
create a state milk-marketing order to force Firm P into pooling Class I and Class II
premiums. After eight months without premium pooling between cooperatives and Firm
P, and with a new state milk marketing order possibly nearing approval, Firm P returned
to the PEC in May 1998. In the new agreement, the Market Balancing Fund was
eliminated and the pooling mechanism modiﬁed from the 1992 agreement. The details
and effects of the changes made as a result of the 1997 price war are also reviewed in the

following section.

54

5.2 The PEC Agreement for Pooling and Transferring Over-Order Premiums

Pooling and transferring over-order premiums to PEC members is not a trivial
task. This section summarizes the operation and mechanics of each of the PEC
agreements providing a technical review of the pooling mechanism and how changes
made after each price war changed the net payoffs to PEC members.

Table 5.3 deﬁnes the variables employed in the following section. Over time, the
PEC utilized six different classes of raw milk sales, and two variable premiums for Class
I and Class II milk. Premiums contributed to the PEC for Class III cheese milk are ﬁxed.
The net results and equations are based on monthly calculations coinciding with monthly
federal order minimum prices. Premiums and prices are on a per hundredweight basis.
Firm j is a generic indicator of market milk marketing organization where j equals
Cooperative B, Cooperative M, or Firm P. Cooperative j is a subset of ﬁrm j, and an
indicator of a generic cooperative milk marketing organization, where j equals

Cooperative B or Cooperative M.

55

Table 5.3 Variables Used in PEC Pooling Mechanism Equations

 

 

Variables Description

Q,’ The quantity of Class I milk sold by ﬁrm j

Q}; The quantity of participating Class II milk sold by ﬁrm j

Qé The quantity of non-participating Class II milk sold by ﬁrm j

QI’), The quantity of participating Class III milk sold by ﬁrm j

Qéiv The quantity of Class III cheese milk sold by cooperative j to a
non-cooperative buyer/processor

Qéc The quantity of Class III cheese milk sold by cooperative j to a
plant partially or wholly owned by a cooperative

Q; The total quantity of milk sold by ﬁrm j for all classes

QiP The quantity of Class 1' milk sold by Firm P

QM The quantity of Class i milk sold by Cooperative M

Q? The quantity of Class 1' milk sold by Cooperative B

Q,C The total quantity of Class 1’ milk sold by PEC cooperatives

P1 The PEC Class I premium

p 1/ The PEC Class 11 premium for participating Class II milk

P31 The PEC Blend Premium

 

5.2.1 The 1985 PEC Agreement

The 1985 PEC agreement is the simplest and most straightforward agreement
analyzed here. In the era governed by the 1985 agreement, Cooperative M and
Cooperative B were the only members of the PEC. Firm P participated in the market, but
did not pool premiums collected from the sale of Class I or Class II milk. This pooling
mechanism governed the PEC ﬁ'om the conclusion of the 1982 price war in 1985, until
the 1992 price war. A similar pooling mechanism operated in the PEC from the time of
the ﬁrst over-order pool, established in Detroit in 1956, until the 1982 price war.

For collecting and pooling premium revenues, the 1985 PEC agreement used four

classes of milk sales, Class I, Q, , participating Class 11, Q

I, , non—participating Class 11,

Q5 , and Class 111, Q,,,. Participating Class II milk was used in the production of

56

traditional soft products such as yogurt and ice cream. Non-participating Class II milk
was used in Class II products sold on a national market, such as condensed cream, so
over-order premiums could not be collected on this milk.

PEC members individually collected the announced Class I and Class H PEC
premiums, p, , and p ,, , on the sale of milk for Class I and participating Class II uses on

a per hundredweight basis. PEC members then transferred all premiums to the PEC and
received a settlement, or net payoff.
The following equations track the ﬂow of payments into and out of the PEC using

the 1985 pooling mechanism. The PEC received all premiums collected by PEC

members, ,685 , on the sale of cooperative Class I milk, Qf , and participating cooperative
Class II milk, Q5, so:

(5.1) a. = Qf(p,)+Q§(p,, ).
The PEC calculated a blend premium, p 3135 , by dividing total premiums by total
cooperative milk, which included non-participating Class H milk, Q: , and Class HI milk,

Q5, , or:

 

ﬁes
(5.2) p = .
Qf +95 +9; +95,

The net payoff Cooperative j received from the PEC, F815, was the product of the blend
premium and total milk sales by Cooperative j, Q; or:
(5.3) F8], = (p 3135 XQ; ), for j = Cooperative B, Cooperative M.

However, (5.3) merely explains the distribution of PEC pooled premiums, not all

the revenue generating activities of cooperatives between 1985 and 1992. As described

57

earlier in this chapter, Cooperative B initiated the 1992 price war because Cooperative M
collected—and did not pool—premiums on Class HI cheese milk. Although these
premiums were not transferred to the PEC, the blend premium calculations included this
milk as a part ofClass III milk.

To incorporate this additional revenue for Cooperative M, these premiums are

added to (5.3) to calculate a new net payoff to Cooperative M, N354 , as:

(5.4) N3: = ((23V XMCN )+ ((29% XMCC)+ 11$,
where MCN represents the premium collected by Cooperative M on the sale of Class III

cheese milk to a non-cooperative processor, Q31, , and MCC represents the premium
collected by Cooperative M on cheese milk sold to a plant partially owned by a
cooperative, Q52. Because Cooperative B did not collect premiums on Class III cheese
milk, Cooperative B’s net payoff equals (5.3).

(5.3) and (5.4) only present the net payoffs for PEC members, speciﬁcally
Cooperative B and Cooperative M. Firm P participated in the market, but did not pool
over-order premiums generated on the sale of Class I and Class II milk. Thus, Firm P’s

net market payoff from collecting over-order premiums from 1985 to 1992, N; , was:

QIPXPI)+(Q; L311)-

(55) N5. =(

This research assumes that Firm P collected identical premiums on the sale of Class I and
Class II milk as market cooperatives or PEC members. Altemately, Firm P could charge

slightly smaller premiums and take market share away from the cooperatives. The latter

practice most likely occurred as Firm P’s Class I market share increased in the early

19908 when it did not participate in the PEC.

58

The pooling mechanism of the 1985 PEC agreement clearly favored Firm P and to
a lesser extent Cooperative M. Firm P, with a high Class I utilization, was able to free-
ride on the PEC and collect premiums on Class I and Class II milk without contributing
to market-balancing services provided by market cooperatives. Cooperative M beneﬁted
as it continued to collect, but not pool premiums on Class III cheese milk. This
additional source of revenue increased the net payoff to Cooperative M, illustrated in
(5.4), by about 1 cent per hundredweight, or approximately $30,000 per month.

Figure 5.1 illustrates net PEC payoffs by ﬁrm on a per hundredweight basis in
terms of December 1998 dollars, employing the pooling mechanism of the 1985 PEC
agreement. The net payoff by ﬁrm is calculated by dividing a ﬁrm’s net payoff from the
PEC by total milk controlled by that ﬁrm. This ﬁgure assumes that the 1985 pooling
mechanism spanned the entire length of the time analyzed in this research, speciﬁcally
from June 1985 through December 1998. This ﬁgure illustrates the equity position of
each ﬁrm. Latter ﬁgures allow comparison of the PEC agreements by ﬁrm. The periods
of the 1992 and 1997 price war are discontinuous in the ﬁgure. Net payoffs to
Cooperative B and Cooperative M are combined in Figure 5.1 and subsequent ﬁgures due
to the closeness of their individual payoff. Typically, their net payoffs only differ by one

or two cents on a per hundredweight basis.

59

Figure 5.1 PEC Firm Payoffs Under the 1985 Agreement in Dec. 1998 Terms

 

 

Cooperative M/B

 

[ ....... FirmP

 

 

2.50 —

2.00 _ :3, :1 5

u
I. I’.’.
. .

1.50 -+

1.00 ~

Payoff ($/cwt)

 

0.50 —

 

 

 

 

 

 

0.00 F‘— —f‘ —’"I ——" "TI—"ﬁv‘V—j "” Ti ‘4' l # TAT" !‘ 'T T""— ' —F "—T T“ ‘— J' A " ’fh— #T 7

I985 I987 I989 I991 I993 I995 I997

 

 

 

5.2.2 The 1992 PEC Agreement

As explained above, the 1992 price war resulted in a PEC agreement with three
pools, Pool A, Pool B, and the Market Balancing Fund (MBF), and three participating
ﬁrms, Firm P, Cooperative B, and Cooperative M. Also, the 1992 agreement allowed
PEC members to retain a portion of their collected over-order premiums.

All PEC members contributed to Pool A and the MBF. Firm P received its PEC
payoff ﬁ'om Pool A, with the Pool A residual, or what remained from total PEC revenues
after Firm P’s settlement, transferred to Pool B. Cooperatives contributed additional
premiums to Pool B and received a settlement from this pool. Contributions to the MBF
were divided over milk balanced in a cooperative plant.

The 1992 agreement formalized six classes of milk, Class 1, Class H participating,

Class II non-participating, Class III participating, Class III cheese milk sold to a non-

6O

cooperative processor, QCN , and Class HI cheese milk sold to a plant partially owned by
a cooperative, QCC . Additionally, cooperatives reported the quantities of milk balanced,

Q, , which included milk used for Class III and some Class II products.

Each month the PEC announced Class I and Class II premiums. PEC members

collected these premiums and transferred all but 30 cents of the Class I premium to Pool

A, and the entire Class H premium to Pool A. Thus, total Pool A revenues, ,69‘2 , equal:

(5.6) 13:; = ZQKp, —0.30)+ 29,30" ),
where 2Q} equals the sum of Class I milk sold by participating PEC ﬁrms, and 2Q},

equals the sum of participating Class H milk sold by participating PEC ﬁrms. The PEC
derived the Pool A blend premium, p 3,92 , by dividing total Pool A premiums by total

PEC milk excluding the two classes of Class III cheese milk, expressed as:

 

.652
(5'7) p A = r' r r r' ’
3192 2Q] +ZQII + 2Q}; + 2Q”!
where 2Q;- equals the sum of nonparticipating Class H milk sold by participating PEC

ﬁrms, and Zqu equals the sum of participating Class HI milk, or all Class HI milk

other than cheese milk, sold by participating PEC ﬁrms.

Firm P’s net PEC payoff from Pool A, Q; , equals the product of the Pool A
blend premium and Firm P’s total milk sales in all classes, Q; , or:

(5.8) 9;; = pgmtgl’ ).

Once Firm P received its payoff, the PEC calculated the Pool A residual, A; .

The Pool A residual is the difference between total PEC revenues and Firm P’s net

61

payoff, or the balance of Pool A after Firm P collected its settlement. Speciﬁcally, the

Pool A residual equals:

(5.9) A; = I39: -952,
The Pool A residual was transferred to Pool B in which only cooperatives participated.
In Pool B, each cooperative contributed an additional 10 cent per hundredweight

premium for Class I milk, a 5 cent per hundredweight premium for Class HI cheese milk

sold to a plant partially owned by a cooperative, and 15 cents per hundredweight for
Class III cheese milk sold to a non-cooperative buyer. Thus, total Pool B premiums, [3,2 ,
are:

(5.10) a: = A3: +(Qf )(0-10)+(QSNX0-15)+(Q§c)(0.05).
where Q,C is the total Class I milk sold by participating PEC cooperatives, Q3, is the
total cooperative Class III cheese milk sold to a non-cooperative buyer by participating
PEC cooperatives, and QCCC is the total Class III cheese milk sold by participating PEC
cooperatives to a plant at least partially owned by a cooperative.

As with Pool A, the PEC derived a Pool B blend premium, pg,2 , by dividing

total Pool B premiums by total cooperative milk in all six classes, or:

ﬁg?

(5.11) p8 = .
3"” QIC+Q§+Q§+Q51+Q§N+Q§C

 

Each cooperative collected a payoff ﬁ'om Pool B equaling the product of the Pool

B blend premium and its individual total milk sales. For Cooperative j, the Pool B net

payoff, ‘I’Jz , is:

(5.12) ‘I’Jz = pg,” (Q; ), for j = Cooperative B, Cooperative M.

62

Following the calculation of Pool A and Pool B payoffs, the PEC calculated the
payoffs from the Market Balancing Fund. All PEC members transferred a 10 cent per

hundredweight Class I premium to the MBF. The MBF paid for the cost of the market
balancing services provided by the cooperatives. Total MBF revenues, ﬁg” , equal the
product of total Class I milk and the 10 cent premium, or:

(5.13) a?“ = ZQ;‘(0.10).
The MBF blend premium, p :15; , results from dividing MBF total premiums by the

quantity of milk balanced by all cooperatives, Qf , or:

; 0.10
(5.14) p212: = 29‘. ).
Qb

Each cooperative collected a MBF payoff that equaled the product of the MBF blend

 

premium and the quantity of milk each cooperative balanced. For Cooperative j, where
Q; is the quantity of milk Cooperative j balanced, its MBF payoff, Ag, , was:

(5.15) A1,, = p33: (Q; l, for j = Cooperative B, Cooperative M.

The total net result from PEC participation in the period ruled by the 1992 pooling
mechanism can now be derived. For Firm P, revenue sources included premiums ﬁom

the sale of Class I and participating Class H milk, plus its PEC payoff. However, Firm P

paid the PEC all but 20 cents of the Class I premium, and the entire Class II premium.

Thus, the net result to Firm P, N; , is:

(5-16) N52 =(Qfxpl)+(QlIIXpll)+Q:2 -(QIP ID; '0-20)-(Q; X1011),

which simpliﬁes to:

(5.17) N5, =(QfX0.20)+ (25,.

63

The net result to the cooperatives is more complex due to additional premiums
collected and pools employed. Cooperative j collected a ﬁxed PEC premium for Class I
milk, Class 11 participating milk, and both classes of Class III cheese milk. Cooperative j
transferred all of these premiums except 10 cents of the Class I premium to the PEC.

Cooperative j received ﬁom the PEC a net payoff from Pool B and the MBF. The ﬁnal

net result, N42 , is expressed as:

(5 18) N42 =(QIJ XPI)+(Q1’)XP11)+(Qé~xo~15)+(Qc’}cx0.05)+ ‘I’g’,
. + No - (Q/ Xp, - 0-10)- (Q) Xpu )- (QéN X0.15)— (Qéc X005),

which simpliﬁes to:

(5.19) N4, = (Q/ XO.10)+ trig, + A1,, for j = Cooperative B, Cooperative M.

The ﬁnal outcomes presented in (5.17) and (5.19) are exhibited in Figure 5.2.
Figure 5.2 shows the net payoffs to Firm P and the PEC cooperatives using the pooling
mechanism of the 1992 PEC agreement on a per hundredweight basis in terms of
December 1998 dollars. Again, the ﬁgure employs the 1992 pooling mechanism for the
research duration and periods of price wars remain discontinuous. In comparison to
Figure 5.1, Firm P payoffs dramatically decrease as Firm P begins to pool Class I and
Class H over-order premiums. Conversely, net payoffs to the cooperatives slightly
increase as they are compensated for balancing services. Additionally, although not seen
in the ﬁgure, net payoffs on average to Cooperative B increase more than payoffs to
Cooperative M, resulting in Cooperative B receiving a higher net payoff than
Cooperative M. These changes occur as PEC participants retain, and do not pool, a

portion of their Class I premiums which favors Cooperative B.

64

Figure 5.2 PEC Firm Payoffs Under the 1992 Agreement in Dec. 1998 Terms

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

....... Firm P Cooperaive M/B
2.50 ,
2.00 a .
g :
o I:
a, 1.50 1 -’.
V I.
to O
3; 1.00 -— ,
a. w:
on o- 9 0‘ ’ ' 0"
0.50 n '
0000 ##""‘_T" *"'_T ’ ﬁ'“—“ T'""' "—T“ _ l 'i T T
1985 1987 I989 I991 I993 I995 1997

Year

 

 

 

5.2.3 The 1997 PEC Agreement

The 1997 PEC agreement modiﬁed the 1992 agreement by eliminating the MBF
and altering the transfer of the premiums between PEC members. Speciﬁcally, PEC
members retain 25 cents per hundredweight of the Class I premium to Pool A, rather than
retaining 20 cents as in 1992 agreement. The remainder of premium revenue ﬂows into
Pool A the same as the previous agreement. In Pool B, cooperatives contribute the same
ﬁxed premiums, but receive a new 15 cent return on Class 11 participating milk from Pool

B.

The net results for the 1997 agreement are similar to the 1992 results as the

pooling mechanism did not signiﬁcantly change. Thus, the net result to Firm P, N5, ,

equals:

(520) N57 = (QIP X10: )+ (Q; X10” )‘l' 9:2 "(91’9pr - 0-25)- (Q5Xp11)’

65

which simpliﬁes to:
(5.21) N5, = (Qf Xo.2s)+ (25,,

where Q; , Firm P’s Pool A settlement, equals:

LELQI’ (PI _ 0:25)}le Qi/ (pl! )]
2Q; + ZQh + 2Q; + ZQIII

For Firm P, compared to the 1992 agreement, the 1997 Pool A settlement, (2; , is

 

(5.22) a; =[ (Q5)-

less than the 1992 Pool A settlement, (2; , as the 1997 net payoff, N; , is greater than

the 1992 net payoff, N52. Using (5.7) and (5.22), comparing the Pool A settlements

results in:

 

(5.23) o; -9; _ ZQK— 0.05)

’ 29; +29}, +ZQ;+ZQ;,,

Similarly, to compare 1992 and 1997 Firm P net payoffs, using (5.17), (5.21), and (5.23),

(Q?)-

the increased revenues to Firm P equals:

 

P P P 2Q; (— 0‘05) P
( ’ (Q I ”29; +29» +29; +29», (Q )

The cooperative net payoff from the 1997 agreement for Cooperative j is:

(5 25) N37 = (91] 1100+ (QM/911% (QéN X0.15)+ (Qgc X0.05)+ 4:917 _
. (Q; X’o’ —0.15)_(Q;,' 1”” ‘0-15l—(Qév X0-15l-(Qéc X005)

which simpliﬁes to:
(5.26) N4, = (Q{ 10.1 5)+ (Q; XO.15)+ trig” for j = Cooperative B, Cooperative M.

The Pool B payoff to Cooperative j, ‘13, , is:

66

 

+(Qf )(0 10)+(Q,€ X—O 15)+(Q§~)(0 15)+ (QECXO .Qos)]( 1).
Q1 + Q11 + Q15 + Q111+ QCN + QCC

(5.27) ‘I’g’, =[A97

In comparison to the 1992 agreement, PEC cooperatives retain a greater portion
of initial Class I and Class II premiums under the 1997 agreement. However, Pool B
payments are less in the 1997 agreement, and the MBF source of cooperative revenue is
eliminated.

The overall result of the 1997 price war was another shift in distribution of PEC
rents at the expense of Cooperative M. The net payoff to Firm P increased, as did the
payoff to Cooperative B. The reason for this increase is twofold. First, PEC members
retain a greater share of Class I and Class II premiums, favoring Firm P and Cooperative
B both with high Class I utilization compared to Cooperative M. Additionally, Firm P
and Cooperative B no longer pay for market balancing through the MBF. Cooperative B
gains through this change as in 1997 it exited balancing. Using the same basis as in
Figures 5.1 and 5 .2, Figure 5.3 shows the net payoff to market ﬁrms using the 1997

pooling mechanism

67

Figure 5.3 PEC Firm Payoffs Under the 1997 Agreement in Dec. 1998 Terms

 

 

 

 

 

 

 

   

 

 

 

 

 

 

2.50 . ------- an P Cooperarve M/B
2.00 . I
A 5.
E "' :
(>9 1.50 2'. ;
V ' I
“a o
5‘ 1.00 - _:.
O- ,1
0.50
0.“) “'—“ ' T—j— ‘ "" TF— -T—' ' _'—T——_'_T— ' __ I‘ I ‘ f I T I
1985 1987 1989 1991 1993 1995 1997
Year

 

 

 

Figure 5.4 shows the actual PEC ﬁrm payoffs using the net results derived above.
Again, net payoffs are converted to a per hundredweight basis and expressed in constant
December 1998 dollars. The ﬁgure best illustrates the difference between cooperative
and Firm P payoffs over time. This gap is greatest when Firm P does not participate in
the PEC between 1985 and the 1992 price war. During this period, Firm P’s average
Class 1 premium is 55 cents a hundredweight greater than the PEC cooperatives. Firm P
payoffs signiﬁcantly decrease when they join the PEC in 1992. The difference between
Firm P and cooperative payoffs increase after the 1997 price war. This is better seen
comparing Figure 5.4 with Figure 5.3. Additionally, Figure 5.4 shows that the typical
payoff to cooperatives generally ranges between 50 and 75 cents per hundredweight.

Over this entire period, Firm P’s PEC payoff is on average 39 cents greater than the

cooperatives.

68

Figure 5.4 PEC Firm Payoffs Under Actual Agreements in Dec. 1998 Terms

 

 

 

 

 

 

 

 

 

 

 

2,50 . E l ------- FirmP Cooperative M/Bj

2.00 - 5;.
A :_ - .
E :« .
O i 3 ‘: ' Z; :
a 1.50 : 2 z: : \.I. ‘0‘ .:
v ' . " . ‘s: ' :0 .
‘5
>1
3: 1.00

0.50 «

0.00 A» -#--T— ,— .--p—-_.e-._-..-,_.....f_ ._.#r__

1985 1987 1989 1991
Year

 

 

 

To illustrate the net effects of the various PEC agreements, Table 5.4 reports net
payoffs on a per hundredweight basis under the three PEC pooling mechanisms. The
different payoffs are calculated under each agreement using quantities from the month of
May for even years between 1986 and 1998, and are expressed in constant December
1998 dollars on a per hundredweight basis. Firm P net payoffs decrease from the 1985
agreement to the 1992 agreement when proprietary ﬁrms join the PEC, and then increases
as a result of the 1997 agreement. Cooperative B payoffs increase ﬁom 1985 to 1992
with PEC participation by Firm P, then increase again under the 1997 agreement as it no
longer pays for balancing and retains a greater portion of collected premiums.
Cooperative B’s payoffs ﬁ'om 1992 to 1997 increase the greatest in 1998, since 1998 it
did not assist in balancing the market. Cooperative M payoffs increase from 1985 to

1992 as Firm P joins the PEC, but decrease under the 1997 agreement. Cooperative M

69

payoffs under the 1985 agreement include ﬁxed, non-pooled cheese milk premiums of 5
and 15 cents per hundredweight for cooperative and non-cooperative cheese milk sales.

Table 5.4 PEC Firm Net Payoffs Per Hundredweight Calculated Under Different
PEC Agreements for May Quantities in December 1998 Terms

5.4a Firm P Net Pajoffs ($lcwt)

 

 

1985 1992 1997
Year Agreement Agreement Agreement
1986 0.95 0.64 0.71
1988 1.19 0.82 0.89
1990 1.64 1.11 1.18
1992 1.48 0.97 1.04
1994 1.26 0.84 0.92
1996 0.88 0.64 0.72
1998 1.24 0.85 0.93

 

5.4b Cooperative B Net Payoffs ($/cwt)

 

 

1985 1992 1997
Year Agreement Agreement Agreement
1986 0.43 0.46 0.48
1988 0.57 0.60 0.63
1990 0.79 0.84 0.86
1992 0.67 0.73 0.75
1994 0.58 0.64 0.65
1996 0.48 0.52 0.54
1998 0.59 0.65 0.71

 

5.4c Cooperative M Net Payoffs ($/cwt)

 

 

1985 1992 1997
Year Agreement Agreement Agreement
1986 0.44 0.45 0.44
1988 0.58 0.60 0.59
1990 0.80 0.84 0.83
1992 0.69 0.72 0.71
1994 0.59 0.63 0.61
1996 0.48 0.52 0.49
1998 0.60 0.68 0.65

 

The balance of this research uses the framework and mechanics of the PEC

presented in this chapter to explain observed ﬁrm behavior in the Michigan milk market.

70

A brief review of industrial organization literature examines price wars and cartel

stability in the next chapter.

71

Chapter 6: Cartels, Price Wars, and the PEC

Previous chapters assessed the institutions and policies of milk marketing as well
as the history and operation of the PEC. Chapter 5 speciﬁcally addressed the three PEC
price wars since 1980. Because this research aims to determine the conditions of stability
and explain the conditions leading to these price wars, this chapter reviews applicable
industrial organization literature. Speciﬁcally, this chapter applies the theories of cartel
stability and price wars to the PEC. The ﬁrst section reviews cartel theory including ﬁrm
conduct, enforcement, and monopoly power. The second section assesses the three
conunon types of price wars and their causes. Each section evaluates the PEC in terms of
this literature, providing the foundation for the market model in following chapters.
6.1 Cartel Theory

In any market ﬁrms have an incentive to coordimte conduct to increase collective
and individual proﬁts. Firms increase price or restrict production through either tacit
collusion or cartel organizations. Tacit collusion occurs when ﬁrms coordinate their
actions despite the lack of an explicit agreement or formal communication. Cartels on
the other hand, are economic organizations consisting of independent ﬁrms that explicitly
agree to coordinate pricing or production activities (Carlton and Perloff). Clearly, tacitly
colluding ﬁrms do not meet the deﬁnition of a cartel, as tacit implies informal
coordination, rather than explicit coordination.

Because dairy marketing agencies-in-common coordinate the pricing activities of
cooperatives and independent farmers involved in ﬂuid-milk markets, marketing
agencies-in-common operate as an economic cartels. However, many people mistakenly

negatively associate the term cartel—because of abuse in the popular press—with a

72

grouping of ﬁrms that coordinates pricing or production activities resulting in harm to
consumers. This thesis does not address consumer effects in this research, but instead
concentrates on ﬁrm conduct.

Using these deﬁnitions in this thesis, the PEC is a price-ﬁxing cartel which
redistributes rents between its members. This section examines the conduct, success, and
problems cartels face as they attempt to increase proﬁts.

6.1.1 Cartel Conduct

Once a cartel forms, two approaches exist for a cartel to coordinate industry
activities and increase proﬁts. Member ﬁrms either agree to restrict output or ﬁx prices.
In both cases, the market demand curve determines either the price or quantity variable
not explicitly controlled by the cartel.

When a cartel does not control entire industry output, it belnves as a dominant
ﬁrm that encounters a competitive fringe consisting of smaller ﬁrms. Given that the
ﬁ'inge participates in the market, the cartel selects the price or production level on its
residual demand curve where marginal revenue equals marginal cost to maximize proﬁts.
The gradual entry of fringe ﬁrms over time may destroy cartel market power in the long-
run (Carlton and Perloﬁ). Thus, assuming cartel members cooperate and do not deviate
from the cartel agreement, cartels most easily sustain long-rim market power in industries
with a small ﬁ'inge where new entry is difficult or impossible.

When cartels follow dominant ﬁrm conduct, market results parallel the dominant
ﬁrm model. Assuming no entry, a cartel generates proﬁts by decreasing output and

increasing price compared to the competitive outcome. However, if entry and growth of

73

this fringe is possible, long-run cartel proﬁts tend toward zero and market price and
output tend to the competitive equilibrium.

Turning ﬁom cartel conduct to individual ﬁrm conduct, regardless of how a cartel
coordinates market activity, ﬁrms face an incentive to deviate from the cartel agreement
to increase individual proﬁts. If a cartel sets prices, by undercutting the cartel price, an
individual ﬁrm increases sales and proﬁts by stealing customers ﬁom other cartel
members. In this scenario, without response from the cartel or other cartel members,
only capacity limitations constrain the deviating ﬁrm from capturing the entire market.
Similarly, if a cartel dictates output, by exceeding its allotted market share and selling
greater quantities at the higher cartel price an individual ﬁrm may increase proﬁts.

Firms often believe the cartel can not identify a deviating ﬁrm to successfully
impose economic sanctions. If all cartel members have similar beliefs regarding cartel
monitoring and enforcement, that cartel fails as members face an incentive to cheat.
Economic literature recognizes that due to this payoff situation more cartels fail than
succeed (e.g., Scherer; Stigler). Because of these incentives, cartel success depends on
the ability to enforce cartel agreements to monitor pricing or production decisions by
member ﬁrms while punishing those ﬁrms who deviate from the cartel agreement
(Carlton and Perloﬂ).

6.1.2 Cartel Enforcement

Cartel policies and mechanisms to prevent cheating and enforce agreements are

necessary to stabilize otherwise ﬁagile cartels. Factors that facilitate cartels, or

facilitating meclmnisms, are inherent industry characteristics that contribute to successﬁrl

74

cartels. In industries lacking these factors, cartels implement enforcement policies to
sustain higher proﬁts.

In some industries, cartel instability stems from industry structure. For instance,
industries with many ﬁrms, or few ﬁrms that market differentiated products, have
difficulty in organizing successful cartels. In other industries, facilitating mechanisms
assist cartels in overcoming these structural challenges by making cartel monitoring and
enforcement easier. For instance, monitoring cartel members is less difﬁcult if there are
few ﬁrms in the cartel, usually less than 10, prices are widely known and do not ﬂuctuate
independently from cartel actions, and cartel members sell identical products at the same
point in supply chain. Public availability of ﬁrm level price and quantity information, or
an industry with perfect and complete information simpliﬁes cartel enforcement as well.
Conversely, cartel monitoring is difﬁcult if many ﬁrms participate in the cartel, prices are
volatile, cartel members sell differentiated products, the market experiences unexpected
shifts in demand or costs, or if price and quantity information is not widely known
(Carlton and Perloﬁ).

Cartels can also implement policies to prevent cheating by member ﬁrms through
devising and implementing punishment strategies. A ‘trigger’ price strategy is one
method cartels employ where a predetermined signal—often a lower market price—
triggers prescribed conduct by cartel members. The prescribed conduct results in a
period of reduced industry proﬁts. This reversion, commonly a reversion to the
competitive equilibrium, occurs regardless of the cause of signal. For instance, if price
decreases because of cheating, the deviating ﬁrm is punished as are all other ﬁrms by

lower proﬁts.

75

A trigger strategy is useful when cartel members lack complete information to
determine the cause of a price decrease. A trigger strategy serves to prevent future
cheating as cartel members know in advance the cartel response to a price decrease.
Individual ﬁrms weigh expected beneﬁts ﬁ'om cheating against the expected losses from
the cartel response. Ideally, the cartel response inﬂicts expected losses greater than the
expected beneﬁts ﬁom cheating, preventing cheating in future periods.

Cartels that operate in industries where cartel members have little incentive to
cheat do not worry about monitoring and enforcement. Firms with little incentive to
cheat exist in industries where marginal cost curves are inelastic, ﬁxed costs are low in
relation to total costs, or when a single sales agent or organization coordinates pricing
activities. A single sales organization effectively provides perfect information to cartels
as members openly share price and quantity information. Cartels whose members have
little incentive to cheat are more successﬁrl in effecting price and operate for greater
lengths of time (Carlton and Perloﬁ).

Organization costs may also affect cartel enforcement and stability. Low
organization costs support cartel formation while high costs prevent cartels from
organizing. Organization costs are low when few ﬁrms participate in the cartel, the
industry is highly concentrated, or the cartel deals with a homogeneous product (Carlton
and Perloff). These same characteristics also facilitate cartel monitoring and
enforcement, as the opposite factors require complex mechanisms to monitor ﬁrm

conduct.

76

6.1.3 Monopoly Power

In addition to adherence to and enforcement of the cartel agreement, to generate
rents cartels must possess monopoly power. Monopoly power, or the ability of a ﬁrm to
aﬁ'ect price, is a function of market control and demand characteristics (Pindyck and
Rubinfeld).

The determination of monopoly power depends on the cartel’s residual demand
curve being sufﬁciently inelastic as well as controlling a majority of the market. Cartels
meet this criteria either through selling a product with an inelastic demand, or the fringe
having an inelastic supply (Carlton and Perloff). The Organization of Petroleum
Exporting Countries (OPEC) provides an example of the latter requirement, as countries
involved in OPEC face an inelastic fringe supply. Despite major price increases, non-
OPEC countries or fringe countries, have not signiﬁcantly increased production in the
short or medium-run (Carlton and Perloﬁ). Conversely, an international copper cartel
faced an elastic supply due to secondary supply markets from scrap. This elastic supply
prevented the cartel from raising prices in the long-run and contributed to the cartels’
failure (Carlton and Perloﬁ‘).

A close substitute available outside the cartel also increases supply elasticity,
whereas potential entry increases demand elasticity. Industries without the ability to
maintain inventories (e. g. perishable products) face decreased demand elasticity which
facilitates cartel formation

Attaining monopoly power by affecting price complements adherence to the cartel
agreement facilitating cartel success. Cartels cannot increase price without satisfying

both requirements. For example, in Kuala Lumpur, a pepper cartel set minimum export

77

prices when it controlled 95 percent of the world’s production. Because members
continually undercut the cartel price, the cartel was unsuccessful in the long-run despite
clearly having potential monopoly power (Carlton and Perloff).

6.1.4 Cartel Theory and the PEC

Applying the theories of cartel conduct, cartel enforcement, and monopoly power
to the operation of the PEC highlights the industrial organization aspects relevant to this
analysis. This section addresses the PEC’s choice variable, market entry, cartel
enforcement, organization costs, and monopoly power.

The PEC operates to increase Class I and Class H milk prices in Order 40. Since
Michigan cooperatives pursue policies of open membership and do not limit members’
supply, production decisions remain the choice of individual Michigan dairy farmers.
Therefore, PEC class prices and market demand curves determine class utilization. By
controlling the entire milk supply, the PEC chooses the Class I and Class H prices to
maximize market revenues. However, these prices are constrained by price of milk and
cost of transportation Item the nearest surplus milk market.

PEC proﬁts should induce entry into the Michigan milk market by organizations
that market raw milk. However, barriers exist that make entry difficult and costly.
Entering the market requires control of a milk supply. A ﬁrm can secure a milk supply in
two ways, through their own milk production activities or bidding away established dairy
producers ﬁ'om another ﬁrm

Milk production requires large capital investments in speciﬁc assets, namely cows
and dairy facilities. To change PEC conduct, that is reduce the level of PEC over-order

premiums, a ﬁrm needs to produce large enough quantities of milk to acquire supply

78

contracts with milk processors presently controlled by the PEC. These constraints limit
entry by ﬁrms through their own production, which makes this option impractical.

Alternatively, a ﬁrm may enter the market by organizing a brokerage milk
marketing organization that buys milk from dairy producers and sells bulk quantities to
processors. Again, to affect PEC pricing policies, the ﬁrm needs to control large
quantities of raw milk. Essentially, the new ﬁrm could sell Class I and Class H milk
slightly below the announced PEC price and pay dairy farmers a slightly higher pay price
than the price they currently receive. This economic incentive, which requires a high
Class I utilization, should lure dairy producers to the new milk marketing organization.
However, many Michigan dairy operations have contracted with the same milk marketing
organization for generations, have built equity with cooperatives, and prefer the
additional services cooperatives provide. Thus, past history and loyalty, in addition to
other services established milk marketing organizations provide, may make bidding away
farmers a daunting task.

In fact, the number of milk marketing organizations in Michigan has steadily
declined between May 1980 and December 1998. Through a series of mergers and
acquisitions, over 10 ﬁrms have exited while only one new ﬁrm entered the market
during this period. With these apparent entry barriers, PEC members receive positive
economic proﬁts in the long-run if the cartel agreement is stable.

Little monitoring and enforcement is needed in the PEC as its members have little
incentive to cheat as detection is inevitable. Fluid-milk cartels are peculiar in terms of
cartel theory as they operate in a world of near perfect or complete information. Each

ﬁrm in every federal order must report the quantities of milk sold by class to a market

79

administrator on a monthly basis. Legal and economic penalties result from ﬁrms
providing false information to the market administrator.

The rmrket administrator reports the quantities of milk sold by each cartel
member to an agent hired by the PEC. This agent calculates the total premiums and net
payoffs to each PEC member based on the quantities reported. Deviating from the PEC
agreement requires withholding premiums ﬁom the pool. If a ﬁrm deviates, all cartel
members would know almost immediately. Thus, the PEC is a cartel with a single
organization that provides complete information.

In terms of organization costs, again the Michigan milk industry is conducive to
cartel formation. Few ﬁrms participate in the industry and the cartel deals with a
homogeneous product. The cartel consists of three major milk marketing organizations—
two cooperatives and one major proprietary ﬁrm—and a ﬁinge consisting of smaller
regional milk processing ﬁrms. In the period analyzed, the total number of market
participants is never greater then 10. Milk, as it is priced ﬁ'om the farm, is a
homogeneous product as it is deﬁned as not less than 3.25 percent fat and not less than
8.25 percent milk-solids-not-fat, such as lactose, proteins, and minerals.

Cartels require monopoly power, consisting of market control and inelastic
demand to raise price above competitive levels. These two requirements are sufﬁciently
met in ﬂuid-milk markets. The current PEC controls 99 to 100 percent of ﬂuid
production in Order 40. Demand for Class I and Class II milk is inelastic. Wohlgenant

found a demand elasticity estimate of -0.3 for ﬂuid milk in the relative range.

80

6.2 Price Wars

Price wars occur when a ﬁrm deviates from a cartel agreement, usually resulting
in sharp price decreases and production increases. Speciﬁcally, during price wars in
price-ﬁxing cartels, price tends toward the competitive equilibrium where price equals
marginal cost. During price wars in quantity-setting cartels, quantity expands ﬁ'om the
monopoly output toward the Cour-not equilibrium where output is a function of the
number of ﬁrms within an industry.

These sections review literature on price wars and their causes: ﬁrst, price wars
as a result of demand booms or busts; second, equilibrium price wars; and third,
bargaining price wars. The last section summarizes additional factors of cartel stability
and performance as they relate to the PEC.

6.2.1 Price Wars in Busts and Booms

Industrial organization literature initially suggested price wars occur in response
to low demand (Scherer). Using this explanation, in periods of low demand, cartel
members with shrinking sales due to the industry ‘bust’ deviate from the cartel agreement
to increase their sales in a sluggish industry. Firms in this situation, in order to satisfy
cash ﬂow needs, turn to short-run decision making and place greater value on immediate
gain rather than long-run cartel proﬁt (Scherer). Price wars might also occur in times of
bust in an industry where borrowing is difﬁcult and a dominant ﬁrm participates. In this
situation, a recession holds an opportunity for the dominant ﬁrm to reduce the number of
its competitors by driving them out of the market through a price war (Rotemberg and

Saloner).

81

Another hypothesis is that price wars occur in ‘booms’, or periods of high
demand, when cartel members face the greatest payoff ﬁ'om cheating (Rotemberg and
Saloner). In price-ﬁxing cartels, the beneﬁt to an individual ﬁrm from undercutting the
cartel price and increasing sales is the highest during booms. In this scenario, during a
boom, the deviating ﬁrm captures a larger market share during the period before other
cartel members respond with a similar price decrease. Deviation might be the optimal
strategy even if punishment occurs. If the punishment occurs in the ﬁtture, after demand
returns to a normal leveL the punishment from deviating is possibly less than the proﬁt
ﬁ‘orn cheating in the immediate period of high demand.

To counter the boom incentive, price—ﬁxing cartels have two options. The cartel
either relinquishes its attempt at collusion during booms and reverts to competition, or
settles for an equilibrium price lower than the equilibrium price during normal demand
that yields the highest level of sustainable joint proﬁts. Settling for a lower equilibrium
price rewards cartel members with higher levels of proﬁts than a competitive reversion.
Because the monopoly outcome is not the only equilibrium, cartels may choose a price
that lowers the level of industry joint proﬁts, and in so doing decreases the beneﬁts ﬁ'om
cheating. For these cartels, prices may actually decrease as demand increases so no ﬁrm
has an incentive to cheat (Rotemberg and Saloner).

6.2.2 Equilibrium Price Wars

Rather than price wars characterizing unsuccessful collusion, some economists
contend price wars may facilitate collusion (Lameroux; Porter; Green and Porter).
Lameroux applied this notion to a market with a price leader or dominant ﬁrm and a

competitive fringe. In this model, the dominant ﬁrm and fringe engage in successful

82

collusion while price wars become a mechanism to sustain ﬁrture collusion. When the
dominant ﬁrms’ market share drops below a critical level, that ﬁrm instigates a price war
and reclaims a satisfactory level of market control. As the dominant ﬁrms’ market share
increases, the ﬁinges’ market share decreases, after which all participants revert to
collusive conduct. In this case, the dominant ﬁrm uses the price war as a tool of
communication with the fringe to sustain long-run collusion.

Altemately, price wars may sustain collusion when they are used to maintain the
credibility of the enforcement mechanism. The trigger strategy is one type of cartel
enforcement policy that works in markets that do or do not contend with exogeneous
sources of demand uncertainty (Porter, 1985). In industries without exogenous demand
shocks, cartel members successfully set a price through collusion. When and if a ﬁrm
cheats, indicated by market price sliding below the cartel price, remaining ﬁrms
permanently revert to non-cooperative price levels. However, when the market contends
with exogenous sources of demand uncertainty, market shares become a random function
of price, which is a function of industry demand. In this case, the cartel amends the
trigger strategy so that deviations in ﬁrm market shares trigger the price war enforcement
mechanism. Usually, cartels choose the equilibrium cartel price to guarantee that
member ﬁrms will not deviate from agreed upon prices in cooperative periods. However,
price wars may occur in response to exogenous shocks beyond the cartels’ control, which
induce large discrepancies between expected and actual market shares. Firms engage in
price wars to deter cheating, which in a world of imperfect information they cannot
distinguish from stochastic ﬂuctuations. Firms may realize that demand shocks, not

cheating, cause the deviations, but in order to maintain the enforcement mechanism in

83

future periods, cartel members respond with the predetermined punishment or
competitive reversion. Such actions insure that the enforcement mechanism is a credible
threat.

Actual demand shocks that trigger price wars may be either unforeseen increases
or decreases to industry demand. Green and Porter suggest cartels may implement the
trigger strategy in response to unexpected business cycle downturns. Essentially, ﬁrms
switch to competition because they confuse the prevailing low prices in equilibrium with
cheating on the part of other ﬁrms. These ‘equilibrium’ price wars do not reﬂect a
collapse in collusion, but are instead a crucial element in its stability.

6.2.3 Bargaining Price Wars

The theories to explain price wars reviewed above deal with identifying explicit
market conditions. However, Levenstein suggests bargaining price wars also occur when
a cartel member publicly announces its intention to violate the cartel agreement to gain a
more favorable distribution of cartel proﬁts. These bargaining price wars may not
necessarily be foreshadowed by demand shocks or needed to facilitate the cartel
agreement.

Levenstein provides two sources of motivation for a ﬁrm to initiate a bargaining
price war: either a ﬁrm experiences a cost saving technology shock or a ﬁrm simply
desires to change the cartel rent distribution. The ﬁrst occurs in a stable industry at some
collusive equilibrium. Through an outside event, a cartel member experiences a
technology shock that lowers its marginal costs. This ﬁrm initiates a bargaining price
war and deviates ﬁom the cartel agreement because the previously stable agreement is no

longer an equilibrium. A new equilibrium results as the low cost ﬁrm lowers its prices,

84

expands output, and remaining high cost ﬁrms reduce output. This type of bargaining
price war is a method of credible communication among continuously colluding ﬁrms.

An alternative motivation for a bargaining price war occurs in the same industry
without the technology shock to one ﬁrm. In this case, a ﬁrm realizes there are several
stable collusive equilibria within the cartel, all with different distribution of rents. One
ﬁrm instigates a bargaining price war to force the cartel to negotiate a new equilibrium
with a rent distribution more favorable to them. The price war in this instance does not
signal new information to other participants, but changes ﬁrm expectations that move the
industry to a new equilibrium with a different distribution of cartel rents.

Levenstein suggests such price wars characterized bromine price wars of the late
18008 and early 19003. In particular, the bromine industry suffered three bargaining
price wars, each of which began with a public announcement that one member would
violate the existing agreement with the explicit intention of changing the distribution of
rents. In each case, industry ﬁrms reverted to the competitive equilibrium until the cartel
negotiated a new cartel agreement with a different rent distribution. These price wars
resulted from basic disagreements that inherently plague the operation of cartels and were
more severe than price wars caused by ﬁrm monitoring problems.
6.2.4 Determinants of Price Wars

In addition to the three price war types, other factors contribute to ﬁrm decisions
that lead to price wars. These factors include patience, cost structure, and expectations.

The patience of a ﬁrm affects ﬁrm decision making relative to cartel participation.
Firms continually decide whether to abide by or deviate from the cartel agreement. In

doing so, ﬁrms weigh future cartel punishments against gains from deviating in the

85

present. Firm ‘patience’ affects this decision based on the time preference of owners and
the ﬁrms’ debt structure. For example, family held businesses may make decisions
conducive to long-run lifestyle needs and passing the business between generations. For
these ﬁrms, long-run proﬁts take precedence over short-run proﬁts. Family held, or
closely held ﬁrms maybe more likely to collude and discount immediate proﬁts to
maximize continuous long-run cartel payoffs. Conversely, ﬁrms with dispersed
ownership maybe more likely to deviate ﬁom cartel agreements to increase short-run
proﬁts when dividends and interest payments are a priority. Firm debt structure also
inﬂuences ﬁrm patience regardless of ownership type. High debt ratios and ﬁxed
payments make ﬁrms impatient where sales today are more important than high cartel
prices in the future when a ﬁrm may be bankrupt.

The level of ﬁxed and marginal costs also affects cartel stability. Cartels are
usually successﬁrl in industries with low ﬁxed cost and high marginal costs. High
marginal costs reduce the marginal payoff ﬁ'om cheating and facilitate cartels. Firms
with low ﬁxed costs can easily respond to declines in demand with subsequent declines in
output without facing the threat of bankruptcy. This ability to respond to demand
declines results in stable cartels in such industries.

Finally, ﬁrm expectations and beliefs have a role in cartels and price wars.
Repeated interaction between ﬁrms creates relationships and develops expectations about
competitor conduct. For instance, ﬁrms that historically cooperate usually expect the

same cooperation between ﬁrms in future periods resulting in successful cartels.

86

6.2.5 Price Wars and the PEC

Using the above explanations and factors that contribute to price wars, this section
assesses the PEC in terms of the economic literature regarding price wars.

Of the three types of possible price wars, two potentially apply to the PEC and
one ﬁts observed market events. Since the PEC operates with near complete information,
equilibrium price wars are not applicable.

If PEC price wars resulted from extreme states of demand, speciﬁcally booms or
busts, the cartel response should be consistent over time relative to changes in demand.
That is, if the state of demand caused PEC price wars, then the price wars should occur in
either periods of low demand or high demand, not both. However, the 1992 and 1997
PEC price wars began at different times of the year at opposite states of demand. Milk
consumption exhibits seasonal trends, peaking in the fall and reaching its low in late
spring or early summer while supply also exhibits seasonality peaking in summer and
reaching its low in winter (see Chapter 2). The 1992 price war started in June when
demand was low, as the 1997 price war began in September when demand was high. The
timing of these price wars discounts the state of demand as a cause of the PEC price wars.
Moreover, examining general US. supply and demand conditions, these periods were
typical as the industry did not incur an unobserved demand or supply shock (USDA
Agricultural Statistics).

PEC price wars ﬁt the form of bargaining price wars as each PEC price war began
with a public announcement of a members’ intent to withdrawal from the PEC.

Moreover, each price war resulted in a change to the pooling mechanism causing a new

87

distribution of cartel rents. Therefore, PEC price wars are hypothesized as general
examples of bargaining price wars.

With the PEC price wars as bargaining price wars in mind, this thesis examines
each of the price wars to determine its underlying cause using a game theoretic
framework. Chapter 7 reviews the basic theories of games and its economic application
Chapter 8 utilizes this review and models ﬁrm conduct in the Michigan milk market and
explains the existence of over-order premiums as well as each of the three PEC price

wars.

88

Chapter 7: Non-Cooperative and Cooperative Games

Chapter 6 concluded that the recent PEC price wars were bargaining price wars
rather than price wars caused by demand shocks (i.e., price wars due to booms or busts)
or imperfect monitoring. With available price and quantity information for the period of
May 1985 through December 1998, only 14 months ﬁ'om periods of price wars are
included making econometric estimates unreliable because of few degrees of freedom.
Instead, a game theoretic model explains observed conduct of PEC members over time.
This chapter reviews basic game theory concepts necessary to understand the models of
the Michigan milk market used in Chapter 8. (For further information and application of
game theory, see Bacharach and Gibbons.)

7.1 The Theory of Games

Game theory models determine and predict behavior in multi-person or multi-ﬁrm
decision making scenarios while accounting for interdependence among ﬁrms. That is,
the payoff a decision maker expects to receive is not only a ﬁrnction of their own
decision, but also depends on the decisions of other players within the game. By
modeling this interdependence, game theory relaxes the assumption under perfect
competition that economic agents behave as if their actions and payoffs are completely
independent of competitor’s actions (Staatz).

In games between ﬁrms, each ﬁrm evaluates the expected utility it receives ﬁom a
speciﬁc course of action based on potential competitor decisions. Figure 7.1 illustrates a
simple one-time game between Firm A and Firm B where decisions are simultaneous. In
this game each ﬁrm chooses between two possible courses of action called strategies.

Firm A chooses either strategy A1 or A2, and Firm B chooses either strategy B1 or B2.

89

Expected payoffs are derived from the different combinations of strategies independently
selected by each ﬁrm. This game has four potential solutions, each described by different
payoffs to the participating ﬁrms.

In the payoff matrix of Figure 7.1, the payoff to Firm A is listed ﬁrst in each
quadrant followed by the payoff to Firm B. For instance, if Firm A chooses strategy A1
and Firm B chooses strategy B1, then Firm A’s payoff is 8 and Firm B’s payoff is 6. If
Firm A chooses strategy A2 and Firm B chooses strategy B1, then Firm A receives a
payoff of 12 while Firm B receives a payoff of 3.

Figure 7.1 Example Game Matrix
Firm A’s Choices

 

 

 

 

 

Al A2
B1 (8, 6) (12, 3)
Firm B’s Choices
B2 (4, 9) (6. 4)

 

Game theory in various forms applies to many different decision making
situations between ﬁrms. The relevant characteristics include the number of ﬁrms in the
game, the nature of the payoffs, the amount of communication, and potential agreements
permitted between ﬁrms.

Game theory may be applied to almost any ﬁrm decision making situation where
the payoff to one ﬁrm is partially based on the decision of at least one other ﬁrm. A
game formally includes the following four elements:

1. A well deﬁned set of possible strategies for each player.
2. Well deﬁned player preferences among the possible outcomes.
3. Outcomes determined by the strategy choices of the players.

4. Knowledge of the ﬁrst three elements by all players.

90

The ﬁrst and third elements describe the rules of the game as each ﬁrm knows all
possible strategies available at the start of each game. That is, when a game begins a
player does not search for the strategies in their choice set. Additionally, the outcomes of
the game are a ﬁrnction only of the possible player strategies within the game.
Exogenous factors do not inﬂuence potential outcomes.

The second and fourth elements describe player characteristics. Element 2
requires that every player has clear preferences that change in response to changes in the
situation of the game. Element 4 requires complete information, that all players know the
rules of the game and every player’s utility function. This last element limits the general
application of game theory, as it is rare that every player knows all other players’ risk
preferences, ﬁrm debt structtn'e, cost information, and other factors that affect a ﬁrm’s
utility ﬁrnction. However, basic assumptions about the players and the payoff allow
games to model rational ﬁrm decision making.

The most straightforward classiﬁcation of games is based on the number of ﬁrms
in the game. There are games that include two-players, referred as two-player games,
and games that include more than two players, referred to as n-player games. The games
used in this research are two-player games.

The nature of payoffs between ﬁrms further characterizes games. Games are
deﬁned as either zero or non-zero sum games. In a zero-sum game, the sum of the
payoffs across all the players is constant, usually at zero. In non-zero sum games, the
sum of the payoffs across all ﬁrms is not zero or constant. In non-zero sum games,

opportunities may exist for ﬁrms to attain higher payoffs through cooperation. The

91

models of the Michigan milk market are non-zero sum games, as was the example in
Figure 7 . 1.

The amount of communication allowed between ﬁrms also determines the nature
of the game. In cooperative games, ﬁrms communicate about coordinating strategies and
entering into binding commitments regarding these strategies. In non-cooperative games,
legal statutes such as anti-trust laws or high costs prohibit the communication and
interaction found in cooperative games. Further details of cooperative games are
presented later in this chapter.

Once a game is described, the goal of game theory is to determine the rational
strategy for each player. The group of rational strategies for all players is the game’s
solution. To illustrate the process of ﬁnding the solution, assume the game in Figure 7.1
is a two-person, non-zero sum, non-cooperative game. The game from Figure 7.1 is re-
displayed below to assist this discussion

Figure 7.1 (Repeated) Example Game Matrix
Firm A’s Choices

 

 

 

 

 

A1 A2
B1 (8, o) (12, 3)
Firm B’s Choices
B2 (4, 9) (6, 4)

 

The outcome or solution to this game is determined through a two-step process of
identifying the rational strategy for each ﬁrm. To identify the rational strategy for one
ﬁrm, the payoffs to that ﬁrm are analyzed holding the strategy of the other ﬁrm ﬁxed. If
Firm A chooses strategy A1 it receives a payoff of 8 or 4 depending on the strategy Firm
B selects. If Firm A chooses strategy A2, it receives a payoff of 12 or 6 depending on the

strategy Firm B selects. To determine Firm A’s rational strategy, compare Firm A’s

92

payoffs controlling for Firm B’s strategy selection. That is, assuming Firm B chooses
strategy B1, Firm A does better by choosing strategy A2, as 12 is greater than 10. To
indicate this choice for Firm A, the payoff of 12 to Firm A is in bold in the northeast
quadrant of the payoff matrix correlating to the strategy of B1 for Firm B. Using similar
logic, assuming Firm B chooses strategy B2, Firm A does better by choosing strategy A2,
as 6 is greater than 4. Again, the payoff of 6 to Firm A is in bold in the southwest
quadrant correlating to strategy B2 for Firm B. Therefore, regardless of the strategy Firm
B chooses, Firm A always does better by choosing strategy A2, and this strategy is the
rational solution to this game for Firm A.

Using the same process for F irm B, Firm B always does better by choosing BZ, as
9 is greater than 6, and 4 is greater than 3. The southeast quadrant in the ﬁgure contains
two bold payoffs that indicate the ﬁnal solution to this game and the expected payoff to
each ﬁrm.

Consider the case where Firm A and Firm B are duopolists in a market and the
payoff matrix represents ﬁrm proﬁts. Also, assume that strategies A1 and B1 represent
setting a collusive price for the product Firm A and Firm B sells. Conversely, strategies
A2 and B2 represent pricing the product at a competitive rate, or undercutting the
collusive price. In this game, both ﬁrms set price at the competitive level although both
ﬁrms are aware of the higher potential joint proﬁts that exist through collusion. If this
game were a cooperative game and the ﬁrms could communicate and coordinate their
strategy selection, Firm A might choose strategy A1 and Firm B might choose strategy
B 1. However, since the players cannot communicate and coordinate their decisions, the

rational solution is non-optimal in terms of market proﬁts.

93

This example is a ‘static game’ that occurs in one period. As ﬁrm interaction
usually spans more tlmn one period, a simple modiﬁcation to the static game allows
modeling of repeated interaction. The result is a multi-period game that ﬁnds the present
value of a repeated payoff stream to a ﬁrm. By discounting the payoff stream, collusion
in the previous game may be the rational solution. However, without explicit
communication, ﬁrms may suffer short-run losses while coordinating their strategies by
signaling. In practice, one ﬁrm chooses the collusive price to signal its rival of its
preferred long-run behavior. The other ﬁrm, then rationally chooses the collusive price
because if it continues pricing at the competitive level the other ﬁrm will return to this
pricing strategy. As a result, if both ﬁrms exhibit sufﬁcient patience, and the value of
long-run monopoly proﬁts exceed the value of short-run gains, ﬁrms can coordinate
behavior without explicit agreements or formal contacts. This type of game is called a
repeated game. A repeated game that has no end, or is inﬁnite, is a supergame.

7.2 The Basics of Cooperative Games

The non-cooperative game discussed above is not sufﬁcient to model the
relationships of ﬁrms within the Michigan milk market. The Capper-Volstead Act of
1922 allows the formation of marketing agencies-in-common in agricultural markets that
cooperate to set prices and collect rents. The organization of marketing agencies-in-
common requires communication and agreement between members which prevents the
use of non-cooperative games. Following is a review of the theory of cooperative games
and an introduction to the notation employed later to model the Michigan milk market.

Games are cooperative if ﬁrms communicate and enter into binding agreements

prior to the actual playing of the game, or ‘game proper.’ These binding agreements

94

commit participating ﬁrms to a speciﬁc strategy in the game proper, which does not
change after the agreements are made.

Cooperative games are especially useful to model non-zero sum games where
elements of common interest lead to cooperation and higher joint returns to the players
(Bacharach). The realization of these higher returns requires cooperation among ﬁrms
about distributing these returns. In addition, the players must agree on the strategy for
each ﬁrm to carry out in the game proper. If the players do not resolve these issues,
c00peration is unlikely (Staatz).

Within cooperative games, several solutions usually exist which represent
different rational distributions of the joint returns. Although ﬁrms involved in a
cooperative game face many strategy choices, rules of Pareto-optimality and securing
each ﬁrm’s security level signiﬁcantly reduces the number of possibilities. Rational
solutions are Pareto-optimal, as a player does not consider a strategy pair if another
strategy pair exists that returns a larger payoff to one player and at least as much utility to
all other players. Furthermore, a player does not agree to any outcome that results in a
smaller payoff than that player’s ‘security level,’ or payoff it attains acting independently.
If a solution is rational for all players and ﬁrms agree to act collectively in the game
proper, the players form a coalition.

Coalitions in cooperative games are groups or subsets of the complete set of
players within a game. In the model of the Michigan milk market, the complete set of
players, 1, includes Cooperative M, Cooperative B, and Firm P, denoted as I = {M, B, P} ,
where M, B and P indicate the three Michigan milk marketing organizations. If C is a

subset of I, where all members of C are also members of I, then C is a coalition. For

95

instance, if C = {M, B} , then C is a subset and coalition of I denoted C Q I .

Technically, coalitions are free to form between members and may include only one

player of a game, or may include all the players of a game. In the latter case, if

T = {M, B, P} , then T c: I , or T is the proper subset of I, which is also referred to as the

grand coalition.

With coalitions, the outcomes of the games become more complex and varied as
payoffs ﬂow to coalitions or individual players, and these payoffs may be transferred
between players or coalitions following the game proper. For the cooperative games in
the next chapter that model the Michigan milk market, these games are ‘incorporated’
and involve ‘side-payments.’ Incorporated games are cooperative games where the
payoffs of the game proper, or the ‘rents,’ accrue to the individual members of the
coalition, rather than to the coalition itself. In the Michigan milk market, the premiums
accrue to each PEC member, rather than the PEC, as a product of each organizations
sales and the announced over-order premium. However, PEC members redistribute these
rents via side-payments. The mechanism that determines the level of the side-payments
between PEC members is agreed upon before the game is played. The ﬁnal receipts a
player receives after all side-payments are made is a player’s net payoff. For instance,
Cooperative B’s net payoff equals its game proper rents minus any side-payment it makes
to other PEC members. When PEC members negotiate the PEC agreement, they barter
over the mechanism of the side-payments which dictates how mket premiums are
divided and determine every ﬁrm’s net payoff.

After a coalition forms, it behaves as an independent player in a cooperative

game. Therefore, the rules deﬁning rational solutions apply. Speciﬁcally, the solution

96

must be Pareto-optimal and the coalition must receive its security leveL The security
level is the minimum payoff each coalition can guarantee itself if it acts independently.
Assume that v(M) is the security level of the coalition consisting only of Cooperative M,
and v(B) is the security level of the coalition consisting only of Cooperative B. If
Cooperative M and Cooperative B form the coalition C, than the security level of the

coalition C is v(C) 2 WM «1» v(B). The payoff to coalitions meets or exceeds the sum of

independent payoffs. This result characterizes the principle of super-additivity, which
provides incentive for the formation of the grand coalition in most cooperative games.

If a solution is Parteo-optimal and exceeds each coalition’s security level, the
solution is rational and lies within the ‘core,’ the set of rational solutions. Often, more
than one solution lies within the core and sometimes the core is empty if no rational
solution exists. Once the core is deﬁned, the challenge of cooperative game theory is to
identify the solution within the core that the players agree upon.

Nash proposed a solution concept for cooperative games. To do so, Nash
identiﬁed a non-cooperative game underlying the cooperative game. In this way,
bargaining in a cooperative game can be viewed as a non-cooperative game consisting of
three elements for each player. The ﬁrst element, a threat, is an action a ﬁrm carries out
unless other players agree to a player’s demand (Bacharach). In cooperative games,
threats include sanctions imposed by one player on another when bargaining breaks
down In a two player bargaining game with Player A and Player B, the respective
threats are denoted as (A and t 13 resulting in a payoff-pair (h A , ha), the ‘threat-point.’ If a

breakdown in bargaining occurs, Player A receives h A as a result of 13 , and Player B

receives 11,, as a result of IA.

97

The second element within the non-cooperative game is a player’s demand. A
demand is a speciﬁc value that represents a player’s desired net payoff. Demands are
denoted d, and d, and result in the payoff-pair (d,,d,) if d, and d, are compatible.
Demands are compatible if the sum of the demands do not exceed what is available in the
game proper. That is, if cooperation yields a joint return of 15, and the sum of d A and d B
is 21, the demands are incompatible. In bargaining, if players’ demnds are
incompatible, a breakdown in bargaining results and the game ends with each player
implementing its threat.

The third element is the players’ security level, which is the payoff a player
independently receives outside the coalition. Player A’s security level is denoted VA , and
Player B’s security level is v3. If the payoff pair (vA ,vB) , the ‘status quo,’ equals
(h A , hB), the game is a bargaining game. In bargaining games, when bargaining fails the
players receive their security level. In bargaining games only two outcomes are possible,
either (v, ,v3) or (d ”(13). However, in the short-run in general cooperative games, the
threat-point and the status quo are not necessarily equal as players may carry out threats
to punish its rival, even if the threat is irrational in the short-run. Predation is an example
of a threat that results in a threat-point less than a ﬁrm’s security level.

Assuming that threats are temporary and the threat-point equals the status quo in
the long-run, in Nash’s solution the rational outcome is the point within the core that
maximizes the product of the differences between each player’s demand and security

level. (See Nash for this proof.) Nash’s solution ﬁnds the strategies that maximize the
product of (d A — VA ) and (d B — VB) , such that this product is positive and the solution is

Pareto-optimal.

98

 

The challenge to players in cooperative games to arrive at the Nash solution is the
process of forming their demands. The non-cooperative game proceeds as follows:
1. Each player chooses a threat that results if bargaining breaks down.

2. The players inform each other of their threats.

3. Each player independently decides on its demand as a ﬁrnction of all the
threats, or dA(tA ItB) and dB(tB ltA) , and the demands are presented.

4. If the demands are compatible, each player receives its demand; if not, players
implement their threats (Nash).

Problems arise in this solution ﬁamework in Step 3 where the demands are quantiﬁed

based on the perceived value of the payoffs resulting from the threats, (h A , hB) or
(vA,vB). For instance, when forming d, , Player A estimates h, or v, based on t3. If
these estimates of h A or VA are wrong, it leads to d, being incompatible with dB. To
see this, remember that the strategy pair (t A ,tB) results in the status quo outcome of
(v,,v,,). Since the goal of Player A in choosing at, is to maximize ((11,, — v,)(d, — v,)),
if ’3 is wrong, then vB is wrong, which results in an error in the value of (d B — v3) .
Therefore, Player A’s value of d, is incompatible with d B , because either the sum of the
demands exceeds the joint return, or the outcome is not Pareto-optimal. If such an error
occurs, the outcome to the cooperative game is automatically the status quo, (v A , v 3)-

This chapter provides the basic terminology, notation and theory of games
required to complete the analysis of the PEC in the next chapter. Table 7.1 deﬁnes basic

game terminology utilized in Chapter 8. Employing these deﬁnitions and the framework

99

of non-cooperative and cooperative games, this research now turns to modeling conduct

within the Michigan ﬂuid milk market.

Table 7.1 Basic Game Theory Terminology

 

Term Deﬁnition
Non-Cooperative Game Game where communication or agreements between players
is prohibited.

Repeated Game Game modeling repeated interaction between players.
Repeated games of unknown or inﬁnite length between
duopolists may result in rational tacit collusion.

Game Proper The actual playing of the game.

Cooperative Game

Side-Payments

Net Payoff

Threat

Threat-point

Demand

Security Level

Game where ﬁrms communicate and enter into binding
agreements regarding conduct prior to the game proper. If
ﬁrms enter into an agreement, a coalition is formed.

Payments in cooperative games made between players or
coalitions redistributing the initial payoffs of the game proper.

Final receipts a player receives after all side-payments are
made.

An action a ﬁrm carries out in a cooperative game unless
other players agree to its demand.

Resulting payoffs when bargaining breaks down in
cooperative games and all players’ threats are carried out.

Speciﬁc value representing a player’s desired net payoff in
cooperative games.

Payoff a player attains acting independently item a coalition.

100

Chapter 8: Explaining Conduct in the Michigan Fluid Milk Market
Until this point, this analysis has focused on the history and mechanics of the PEC
emphasizing conduct since 1980. Industrial organization literature and game theory
provide a ﬁarnework for ﬁuther analysis. The theories of non-cooperative and
cooperative games introduced in Chapter 7 are employed in this chapter to analyze the
existence of over-order premiums and conduct in the Michigan ﬂuid-milk market.

The format of this analysis is proposition and informal proof. The primary

 

objectives are explaining the existence of over-order premiums, cooperative conduct, and
long-run PEC stability. Because the existence of over-order premiums are not widely
understood, the ﬁrst proposition demonstrates that tacit collusion by cooperatives to
collect over-order premiums is long-run rational. The second proposition examines
conduct tlmt might explain the 1982 price war along with the role and ﬁrm-level effects
of the formal PEC agreement. The third proposition identiﬁes a potential cause of the
1992 price war, and the fourth explains proprietary ﬁrm ‘participation’ in the PEC since
1992.

This analysis is limited to the Class I market because ﬂuid milk generates
approximately 95 percent of the PEC’s monthly over-order premium revenues. The
informal proofs depend on the model of the Michigan milk market structtu'e consisting of
three players, speciﬁcally Cooperative M, Cooperative B, and Firm P.

Proposition 1: Over-order premiums are generated by rational decisions of
cooperatives in the Michigan ﬂuid-milk market.

The illustration of this proposition requires its separation into two smaller

propositions and proofs. The ﬁrst part examines pricing conduct in the Class I market

101

between Cooperative B and Cooperative M and concludes tacit collusion between these
ﬁrms is long-run rational. That is, Michigan cooperatives decide independently to
behave as though they had a formal agreement. The second part modiﬁes the analysis by
including balancing services Cooperative M performs. Although, this addition does not
affect the conclusion of the ﬁrst part, it more accurately reﬂects conduct within the
Michigan milk rmrket.
Proposition 1 a: Michigan cooperatives can generate over-order premiums without a
formal PE C agreement.

Proof: To explain the existence of over-order premiums in the Michigan ﬂuid-milk
market, this proposition examines Class 1 pricing of Michigan cooperatives assuming
they cannot explicitly cooperate and form the PEC. That is, the two Michigan
cooperatives are modeled as rivals in the Class I market. Since the federal orders set a
minimum Class I price, the level of the over-order premium—which each cooperative
selects independently—is the strategic variable in this game.

This game uses monthly 1998 market sales averages for each cooperative that
were reported in Table 3.2. In 1998, the Michigan market consumed on average 174
million pounds of Class I milk monthly. Cooperative B and Cooperative M compete for
approximately 124 million pounds of Class I sales as Firm P controls the residual. Total
milk volume controlled by each cooperative is important to this game as it determines
potential strategies and payoffs. Cooperative M controls 233 million pounds of milk as
Cooperative B controls 68 million pounds of milk. The payoffs in this game are the
products of a cooperative’s Class I sales and its over-order premium. (All quantities and

premiums are expressed on a per hundredweight (cwt) basis unless otherwise noted.)

102

 

This proof relies ﬁrst on a static non-cooperative game and later move to a
repeated non-cooperative game. The cooperatives choose between two strategies, tacitly

collude and collect a Class I over-order premium, p > 0, or act independently and under-
cut the Class I premium by a margin, m, resulting in a net premium equal to p — m

( p 2 m 2 0 ). For simplicity, p represents the largest premium the market will bear. In
addition, depending on p , m can be the deviation resulting in the competitive market
price for Class 1 milk. That is, or may equal p. This game does not address how the
level of p is selected on a monthly basis, rather its concern is the existence of p.

If both cooperatives choose to tacitly collude, the payoff to Cooperative M is the
product of the Class 1 premium and 770,000 cwt of Class I sales, while the payoff to
Cooperative B is the product of the Class 1 premium and 470,000 cwt of Class 1 sales.
These quantities represent how the market divides Class I sales between the cooperatives
when they follow identical pricing strategies. If both cooperatives act independently, the
payoff to each cooperative is the product of these same quantities and the premium

p — m. Potential changes in aggregate ﬂuid demand in this simple static game are

ignored because these quantities are annual averages. Due to the Class I inelasticity
consumption remains relatively constant over time.

Since milk is homogeneous, Class I buyers rationally switch suppliers based
solely on price. (Non-price factors that determine supply agreements are ignored.) When
cooperatives choose divergent price strategies, the cooperative with the lowest price
captures all the market it can supply. For instance, when Cooperative B acts
independently, and Cooperative M acts cooperatively, Cooperative B increases its Class I

sales at a cost to Cooperative M. Speciﬁcally, Cooperative B increases sales ﬁom

103

470,000 cwt to 680,000 cwt, while Cooperative M’s share of the Class I market decreases
ﬁom 770,000 to 560,000 cwt. (This change in market share is based on the sum of the
quantities remaining constant at total consumption of 1.24 million cwt. It ignores
potential changes in membership and therefore production, instead concentrating on the
cooperatives as milk sellers.) The total quantity of milk controlled by Cooperative B,
680,000 cwt, limits Cooperative B’s total Class I sales in this scenario. Conversely, if
Cooperative M acts independently while Cooperative B chooses to cooperate,

Cooperative M supplies the market’s entire Class I demand of 1.24 million cwt of milk as

 

Cooperative M has sufficient volume to supply total Class I demand. (In reality, this
quantity may be less due to long-run supply contracts Cooperative M may enter into with
Class III processors.) In this scenario, Cooperative B is unable to sell milk to the Class I
market and disposes of its supplies to the manufacturing product markets. The Class HI
market is the national residual or surplus market and is always available to absorb excess
milk.

Figure 8.1 illustrates the payoff matrix of this static non-cooperative game. The
strategy indicators ‘C’ and ‘I’ represent ‘tacitly collude’ and ‘independent’ actions, the
decision the cooperatives face. For each strategy pair, the payoff to each cooperative is
found in each quadrant with the payoff to Cooperative M listed ﬁrst followed by the

payoff to Cooperative B.

104

Figure 8.1 Static Non-Cooperative Game In the Class 1 Market between
Cooperative M and Cooperative B
Cooperative M

6C9 ‘19

 

‘C’ (770.000X P );(470.000X P )" (1,240,000X P " m );(0)( P )
Cooperative B

 

‘1’ (560,000)( p );(680,000)( p -— m) (770,000)( p — m );(470,000)( p — m)

 

 

 

 

*(Net Payoff Cooperative M; Net Payoff Cooperative B)

The value of the margin, m, is large enough to induce Class I buyers to switch
suppliers in the short-run but small enough so that independent action generates a greater
payoff than tacit collusion. Combined with the quantities ﬁ'om Figure 8.1, this
assumption results in cooperative payoff relationships.

For Cooperative M:

(8.1) (770,000)(p) < (1,240,000)( p — m), and (560,000)( p) < (770,000)( p - m) ,

and for Cooperative B:

(8.2) (470,000)( p) < (680,000)( p — m), and (0)( p) < (470,000)( p — m).

Identifying the strategy for each player resulting in the highest payoff, holding the
strategy of its opponent ﬁxed, determines the solution in this non-cooperative game. In
this game, there exists a strong incentive for each cooperative to act independently
regardless of its opponent’s strategy. From (8.1), since
(1,240,000)( p — m) > (770,000)( p) and (770,000)(p — m) > (560,000)(p) , Cooperative
M always fares better by acting independently or choosing strategy ‘I.’ Likewise, from

(8.2), Cooperative B always does better by choosing ‘1’ since (680,000)( p — m) >

(470,000)( p) and (470,000)( p — m) > (0)( p).

105

 

Therefore, both cooperatives acting independently is the rational solution to this
static game. This solution is the Nash equilibrium identiﬁed by payoffs in the southeast
quadrant of Figure 8.1. In this case, cooperatives realize these payoffs although both
cooperatives could realize larger payoffs through tacit collusion.

However, static games do not accurately model rational ﬁrm behavior in markets
where players interact repeatedly over time. To allow for this repeated interaction in
games, the static game is transformed into a repeated game. Repeated games capture the
present value of a player’s payoff stream. The present value of a payoff in the future is
the lump sum today that makes the player indifferent between these two values. The
present value of a ﬁrture payoff for Player X, is the product of Player X’ 8 discount factor,

6X , and the expected future payoff, where 5X equals:

1

(8-3) 5X = (1+kx) 9

 

where k X is Player X’s rate of return.
In repeated games, a player receives a payoff in each future time period I. If It, is

a payoff in period t, the present value of a series of future payoffs is expressed as:
(8.4) 7:, + (5an + 6X2”, + = zo'xt-la, .
t=l

If 7:, and 6X are constant over time, and 7:, is received in perpetuity, the discount factor

simpliﬁes to:

1
8.5 5 =—,
( ) X k

X

106

 

and (8.4) simpliﬁes to:

(8.6) Zaf‘n, = -,:’—.
(=1

x

Using (8.6), the static game from Figure 8.1 may be transformed into a repeated
game. A rational solution is found by assuming the net payoffs and discount factor for
both cooperatives remain constant over time.

To model conduct in this repeated game, Cooperative M adopts a ‘trigger
strategy.’ A trigger strategy is when an initiating ﬁrm adopts the collusive price in the
ﬁrst period as a signal to its rival of its willingness to tacitly collude. That is, one ﬁrm
shows its desire to collude by enduring at most a loss in the ﬁrst period until its opponent
adopts a similar strategy at the end of the ﬁrst period. If the opponent does not adopt the
collusive price in the ﬁrst period, or adopts the collusive price in the ﬁrst period but
deviates in a later period, the initiating ﬁrm adopts the competitive price for some
unknown length of time.

In this game, Cooperative M adopts the trigger strategy and chooses ‘C’ in the
ﬁrst period. Cooperative M continues to choose ‘C’ in every period until Cooperative B
fails to collude. Once a strategy pair other than (C, C) results, Cooperative M chooses
strategy ‘I’ for n periods thereafter. From Figure 8.1, Cooperative B’s best response to
C00perative M’s strategy ‘I’ is to also select strategy ‘I.’ However, if both players adopt
the trigger strategy, then the outcome in every period of the inﬁnitely repeated game is
successful tacit collusion, or payoffs found in the northwest corner of the game matrix in
Figure 8.1.

To determine if the trigger strategy is rational for both cooperatives, assuming

Cooperative M adopts the trigger strategy, and this game shows that Cooperative B’s best

107

 

response is also the trigger strategy. To accomplish this, the present values of the payoffs
are needed.

In the ﬁrst period, Cooperative M chooses strategy ‘C’ and Cooperative B faces a
choice between strategy ‘C’ and ‘I.’ If Cooperative B chooses ‘I’ in the ﬁrst period, in
the next 11 periods—a signiﬁcant amount of time not known by Cooperative B—
Cooperative M chooses ‘I’ as well resulting in the strategy pair (1,1). In this scenario,
Cooperative B receives a payoff of (680,000)( p — m) in the ﬁrst period, and

(470,000)( p — m) in each period for 1: periods. Therefore, the present value of

Cooperative B’s payoffs when it acts independently, 7r; , is:

(8.7) 1:], = (680,000)(p — m) + Z (470,000)(p - m)(§;,"‘),

t=l
for n + 1 periods. The ﬁrst term assumes Cooperative B increases Class I sales as shown
in Figure 8.1 from the payoff located in the southwest corner. Altemately, if Cooperative
B adopts the trigger strategy and plays ‘C’ in the ﬁrst period, then Cooperative B’s

payoffs in this scenario, 7r; , are:

(8.8) 7r; = (470.200XP) for n —) oo , or if n is known and ﬁnite,
B

 

1:; = (470,000)( p) + Z (470,000)(p)(5f,“) .

t=l
Assuming n is large, Cooperative B rationally adopts the trigger strategy—and plays ‘C’
in the ﬁrst period—if hf,j > ”1’3. For this inequality to be true, Cooperative B must be
willing to sacriﬁce the short-run payoff from (8.7) for long-run gains. The reverse of this

proof also holds. If Cooperative B adopts the trigger strategy, Cooperative M rationally

adopts the trigger strategy.

108

 

Applied to the Class I market, when both cooperatives adopt the trigger strategy,

the present value of the payoffs are:

 

(3.9) ”g = (470.:ooxp) , and
B

”a = (770,000)(p)

M kM

 

where It; is the present value of payoffs to Cooperative M when it adopts the trigger

strategy. The result is, Irg > It; and hf, > 711’” for large values of n, even without the

PEC or a formal agreement. Thus, Michigan cooperatives rationally tacitly collude and

collect over-order premiums equal to p in every period.

However, this repeated game does not model the balancing functions performed
in the Michigan market. Including this function and cost to Cooperative M’s payoffs
may affect the solution of this game, as in every period Cooperative M faces an incentive
to increase its Class 1 sales and decrease its balancing costs by choosing strategy ‘I.’
Proposition 1b includes the balancing of Cooperative M in this repeated game model.
Proposition 1 b: Balancing functions do not change the rational outcome of tacit

collusion.
Proof: This proposition extends Proposition la to account for the balancing ﬁrnction
performed by Cooperative M. To include balancing responsibilities the game accounts
for the costs, but not revenues, incurred by Cooperative M to internally convert residual
milk, or raw milk not directly sold to the Class 1, Class II, or Class III markets, into
storable manufactured products. This ‘residual cost’ is net of revenues ﬁ'om balancing

the market and is denoted as c(r) , where r is the quantity of milk balanced. Since the

game utilizes 1998 data, balancing only affects Cooperative M, as Cooperative B does

109

 

not own manufacturing facilities. Ignoring ﬁxed costs, assume that as Cooperative M’s
Class I utilization increases, the quantity of milk balanced and thus the cooperative’s
residual cost decreases.

The net of residual balancing activities may be positive depending on the national
markets for manufactured products. This proof assumes these residual activities are only
costs over time as positive proﬁts should attract entry into these types of operations.
However, because the opposite trend has been observed in the Michigan milk market—
Cooperative B sold its balancing plants—this assumption is defensible.

Figure 8.2 illustrates the payoff matrix of the non-cooperative game adding the
residual cost to Cooperative M’s payoffs. In an average month in 1998, Cooperative M
balanced 646,931 cwt of milk when the cooperatives followed similar pricing strategies.
However, r changes depending on the cooperative’s Class I pricing strategies. If
Cooperative M selects strategy ‘I’ and Cooperative B selects strategy ‘C’, the payoff to
Cooperative M in the northeast corner is realized, and the quantity of milk balanced
decreases ﬁom 470,000 cwt to 176,931 cwt, as Cooperative M pulls milk from balancing
to supply the entire Class I market. If Cooperative B selects strategy ‘I’ and Cooperative
M selects strategy ‘C,’ the quantity of milk balanced increases ﬁ‘om 210,000 cwt to
856,931 cwt, as C00perative B increases Class I sales at a cost to Cooperative M. The
bolded payoffs again represent the Nash equilibrium in the static game displayed in

Figure 8.2.

110

 

Figure 8.2 Non-Cooperative Game In the Class I Market between Cooperative M

and Cooperative B Including Balancing
Cooperative M

 

 

 

 

 

 

‘ C a s I a

to: (770,000)( p)—c(646,93l); (1,240,000)(p -m)-€(1769931);

(470,000)( ,0 )" (0)(P)
Cooperative B
‘1’ (560,000)( p )—c(856,93 1); (770,000)( p - m )—c(646,93l);
(680,000)( P - or) (470,000)( P - m)
*(Payoff Cooperative M;
Payoff Cooperative B)

Similar to Proposition 1a, to model actual market conduct, repeated interaction
must be accounted for. To ﬁnd the rational solution of this repeated game, assuming
Cooperative B adopts the trigger strategy this proof shows that the trigger strategy is also
Cooperative M’s best response. Cooperative B adopts the trigger strategy in the ﬁrst
period to signal its willingness to tacitly collude to Cooperative M as the cost of
divergent pricing strategies is low compared to the long-run payoff stream.

To ﬁnd this solution, the present value of Cooperative M’s payoff stream is
utilized. By deﬁnition, Cooperative B selects strategy ‘C,’ Cooperative M then decides
between the strategies ‘I’ and ‘C,’ knowing that if it chooses ‘I,’ for n periods thereafter

the strategy pair (1,1) results. If Cooperative M selects strategy ‘I,’ the present value of its

payoffs, ”(4 , is:

(8.10) 7:1,, = ((1,240,000)( p — m) — c(l76,931))+ 2 ((770,000)( p — m) — c(646,931)X5;' ).

(=1

Altemately, if Cooperative M chooses to cooperate in the ﬁrst period and every period

thereafter, the present value of its payoff, of, , equals:

111

 

(8.11) ac = (770,000)(p ) —c(646,93 1) for n —> oo , or if n is known and ﬁnite,

M kM

 

it; = (770,000)( p) — c(646,93 1) + 2 ((770,000)( p) - c(646,931)X6,I;' ).

t=l
The trigger strategy is rational for Cooperative M if and only if 72f, > 7:1, in n + 1

periods. Because this inequality holds for large n, cooperation on the part of Cooperative
M and Cooperative B in the Class I market is rational, and over-order premiums are
generated without a forrml agreement such as a PEC, even when controlling for market-
wide balancing.

This solution has some unique consequences, as in this game Cooperative B
rationally free-rides on the balancing services provided by Cooperative M. Additionally,
Cooperative M has no direct incentive included in its payoff to provide market-wide
balancing. However, since this service is necessary in the market to preserve orderly
marketing, Cooperative M acts as the ‘missing hero,’ balancing the market without being
adequately compensated. (Previously, federal price support programs made balancing
operations a proﬁtable enterprise for dairy cooperatives resulting in many cooperatives
serving as a missing hero. With the termination of these programs, fewer cooperatives
perform this role leading to milk shortages and surpluses resulting in increased price
volatility.)

This result may seem unreasonable because Cooperative M shoulders the entire
cost of balancing the market. However, this model explains conduct observed in
Michigan during the 1982 price war. Although the price war lasted 32 months,
Cooperative B and Cooperative M tacitly colluded for the ﬁrst 26 without a formal PEC

agreement. Table 8.1 displays the Class I and Class II premiums collected by

112

 

Cooperative M and Cooperative B during 1984, the last 11 months of tacit collusion.
(Available data ﬁ'om limited this table to 1984 premiums only.) In Table 8.1 the
premiums with an asterisk indicate the months of divergent pricing strategies, without
tacit collusion, by Michigan cooperatives.

Table 8.1 Premiums Generated from Tacit Collusion in 1984

 

 

Class 1 Premium ($/cwt) Class 11 Premium ($/cwt)
Month Cooperative M Cooperative B Cooperative M Cooperative B

January 0.67 0.67 0.14 0.14
February 0.81 0.81 0.13 0.13
March 0.81 0.81 0.26 0.26
April 0.80 0.80 0.10 0.10
May 0.78 0.78 0.15 0.15
June 082* 0.79““ 0.17 0.17
July 0.78 0.78 0.1 1 0.1 1

August 0.77 0.77 0.00* 010"

September 0.83 0.83 0.09* 000*
October 0.86 0.86 0.09 0.09
November 0.91 0.91 0.33 0.33
December 0.00 0.00 0.00 0.00

 

The market over-order premium fell to zero in December 1984 when Cooperative
M adopted strategy ‘I,’ or priced its Class I milk at the federal order price. The PEC
reformed in June 1985 and premiums returned to a similar level as in Table 8.1. The
adoption of strategy ‘I’ by Cooperative M is explained in the following proposition.
Proposition 2: The potential of a formal agreement may generate bargaining power for

Cooperative M.

Proof: From Proposition 1, Michigan cooperatives do not require the PEC to extract
over-order premiums. However, since 1980, except for the 26 months of tacit collusion
and 20 months of price wars, the PEC has operated to formalize over-order premiums.

This proposition explains the existence of the PEC as a formal agreement, instead of tacit

113

 

collusion since 1980, its consequences for Cooperative B and Cooperative M, and a
possible reason for the start and ﬁnish of the 1982 price war.

In explaining the existence of over-order premiums in milk markets above, the
ﬁrst proposition ignored the formal PEC agreement. Speciﬁcally, the existence of the
PEC transforms the non-cooperative game from above into a cooperative game.

In cooperative games, players communicate and enter into binding agreements to
form a coalition. In this game the coalition is a marketing agency-in-common, the PEC.
The Capper-Volstead Act also allows marketing agencies-in-common to transfer rents
between members via side-payments.

Applied to the Michigan market, Cooperative M and Cooperative B communicate,
enter into a quasi-binding agreement, and form a marketing agency-in-common. The
PEC agreement is quasi-binding as it allows members to exit with advanced notice. In
the game proper members of the marketing agency-in-common agree to set prices by
determining the Class 1 premium. The game proper ends once each cooperative receives
payment from its buyers for its Class I milk including any over-order premiums. After
the game proper, cooperatives transfer a portion of these total premiums to each other via
side-payments according to the terms of the PEC agreement.

The players in this cooperative game are the two Michigan cooperatives who are
potential marketing agency-in-common or cartel members. The rents a cooperative
receives are the product of its Class I sales and the Class I premium. Cartel rents are the
sum of the rents individually collected by PEC cooperatives, or cartel members. In
incorporated cooperative games, players redistribute their rents via side-payments to

other players belonging to the same coalition. In this incorporated game, Cooperative B

114

redistributes a portion of its rents to Cooperative M via side-payments. The PEC
formalizes and facilitates these side-payments. The net payoff in this game is what each
player retains after all side-payments. Using these deﬁnitions, a two-player incorporated
cooperative game between Cooperative M and Cooperative B explains the role of the
PEC in the Michigan milk market. This game shows that since 1980, PEC operation may
be explained by Cooperative B’s short-run constraints. Speciﬁcally, this game suggests a
possible explanation of Cooperative M’s choice of strategy ‘1’ during the 1982 price war,

as well as Cooperative B’s return to the PEC in 1985.

A cooperative game requires three elements for each player X, a security level vX ,

a threat t X , and a demand d X . The players in this game are the two Michigan

cooperatives and are denoted ‘M’ and ‘B.’ Recall that the security level is the payoff a
player is guaranteed outside the coalition, a threat is a strategy tlmt results in a certain
payoff, h X , if a player’s demand is not met, and a demand is the net payoff a player
desires from a potential coalition.

In this context, cooperative games answer two questions. The ﬁrst is whether or
not a potential coalition forms, the second is once the coalition forms how the coalition’s
rents are distributed among its members. This proposition examines the ﬁrst question
while a later proposition examines the second.

Coalitions form under the constraints that the ﬁnal solution is Pareto-optimal and
each player’s net payoff is equal to or greater than its security level. This proof assumes
Pareto-optimality in the formation of the PEC, and tests if the PEC meets the second
constraint. To do so, the outcomes ﬁom Proposition 1b are employed. Speciﬁcally, from

(8.9) and (8.11) each cooperative’s long-run security level ﬁ'om tacit collusion is:

115

 

(8.12) v3 = 7:; = (470.200XP) , for Cooperative B, and,
B

vM = hf, = (770’000) (1:) _ C(646’93 1) , for Cooperative M.
M

Cooperative j joins the PEC if and only if the present value of its PEC payoff is

greater than v]. From the net equations in Chapter 5, Cooperative B’s PEC net payoff is

less than its initial collected rents, whereas Cooperative’s M PEC net payoff is greater
than its initial collected rents (premiums). This disparity occurs because Cooperative B
makes side-payments to Cooperative M after the game proper, reducing its rents while

increasing Cooperative M’s. This side-payment, S B , is added to (8. 12) which results in

the net payoffs item the PEC to Cooperative B as:

_ B
(8.13) N8 = (470,000c)(p) S , and Cooperative M as,
B

 

_ (770,000)(p) — c(646,931) + S”
kM '

NM

 

The rational rule of coalitions states:

(8.14) PEC= {B,M}, ifand only if: N8 2 v3 and NM 2 vM ,
where PEC denotes the coalition. Using (8.12) and (8.13), N3 < v3 so (8.14) does not
hold. That is, joining the PEC is irrational for Cooperative B.

To explain why Cooperative B joins the PEC requires a look at each cooperative’s

short-run security level and threat. An additional result of Proposition 1b was the
quantiﬁcation of the short-run threats and the threat-point, (hB, hM) . modeled ﬁ'om the

strategy pair (1,1) in the non-cooperative game in Figure 8.2. That is, each cooperative

116

retains the threat to price Class 1 milk at the federal order minimum in any period. If a

cooperative implements this threat, the resulting threat-points are:

(8.15) ha 2 :(470,000)(p — m)(6,'3"), and,

hM = Z ((770,000)(p - m) - c(646,931)Xo‘,‘,;l ),

i=1
where i is the number of periods a cooperative implements its threat. These payoffs do
not allow for the period of divergent pricing witnessed in Proposition 1. Instead, if one
cooperative carries out its threat, this game assumes the other cooperative immediately
implements ‘I’ in the same period.

Using (8.13), and (8.15), a new rule of rational coalitions is determined
comparing the threat-point to the PEC payoff. Because the PEC payoff remains constant
over the short and long-run, the rule of the PEC in the short-run is:

(8.16) PEC= {B,M}, ifand only if: N3 ZhB and NM 2 12M.

The PEC is a rational coalition if Cooperative B prefers the long-run PEC payoffs over
payoffs associated with the threat-point for 1’ periods. From (8.14) the PEC is rational for
Cooperative M.

However, from Proposition 1, strategy ‘I’ is irrational in the long-run for
Cooperative M. With the possibility of the PEC, the distribution of the rents is
negotiable, whereas in the non-cooperative game, the distribution is ﬁxed which favors
Cooperative B. With a potential cartel, strategy ‘1’ is rational in the short-run for
Cooperative M if Cooperative M’s threat is credible and Cooperative B prefers the PEC
payoff to the threat-point payoff for i periods. That is, if Cooperative B is constrained in

the short-run by debt, cash ﬂow needs, or other factors making it prefer short-run to long-

117

run proﬁts, Cooperative M rationally plays strategy ‘I.’ As a result, Cooperative B joins
the PEC to secure some level of premiums in the short-run rather than endure
Cooperative M’s threat, and resulting threat point for 1' periods. However, if Cooperative
B is not short-run constrained, it should endure i periods of irratioml Cooperative M
behavior, knowing in the long-run Cooperative M will behave rationally and tacitly
collude without a formal agreement.

This model coupled with behavioral observation supports the explanation that
Cooperative B is constrained in the short-run, at least more so than Cooperative M. This
time horizon asymmetry results in a formal cartel agreement and the operation of the
PEC. To thoroughly examine the possible constraints of Michigan cooperatives is
beyond the scope of this research.

The cost of Cooperative B’s short-run constraint is a gain to Cooperative M, the
monthly side-payment. Cooperative M improves its bargaining position, secures
compensation for market-wide balancing, and receives a share of Cooperative B’s rents
through the PEC.

This game theory model represents a potential explanation of conduct associated
with the 1982 PEC price war. In October 1982, perhaps realizing payoffs were higher
outside the PEC through long-run rational collusion with Cooperative M, C00perative B
exited the PEC. Because tacit collusion is long-run rational, Cooperative M and
Cooperative B colluded for 26 months. After losing membership and volume to
Cooperative B, in the 27th month, Cooperative M, desiring a return of the formal PEC
agreement with higher payoffs, implemented its threat of competitive behavior of pricing

Class I milk at the federal order minimum plus a small market service charge. Perhaps

118

 

because Cooperative B is short-run constrained and could not endure the threat-point
payoffs for an unknown amount of time, it returned to the PEC six months later.

The PEC operated with Cooperative B and Cooperative M as its members from
June 1985 to June 1992, during which time Cooperative B may have been short-run
constrained. In June 1992, Cooperative B again exited the PEC. Although short-run
constraints might explain the existence of the PEC, they do not explain exits ﬁ'om the
PEC. Proposition 3 analyzes a potential cause of the 1992 price war and factors
contributing to its reorganization in December 1992.

Proposition 3: Incompatible demands may have caused the 1992 price war. While

compatible demands might have caused the re-organization of the PE C
Proof: To explain the 1992 PEC price war requires a cooperative game model assuming
the PEC exists. Speciﬁcally, this proof models the PEC using a two-player cooperative
game, assuming Cooperative B is short-run constrained and prefers to join the PEC rather
than endure Cooperative M’s threat for i periods. Here this thesis reviews conduct of
Michigan cooperatives and identiﬁes a possible cause of the 1992 PEC price war after the
coalition formed.

As with all cooperative games, a range of rational distributions of PEC rents exist
within the ‘core.’ Which distribution is agreed to by the cooperatives depends on
bargaining. If a break down in bargaining occurs, the result is the threat-point (ha, hM ).
In Nash’s solution of cooperative games, one rational outcome exists within the core. As
the short-run threat-point equals each cooperative’s security level, or status quo, this
cooperative game is a bargaining game. In bargaining games, the Nash solution

identiﬁes the solution in the core upon which the players agree.

119

 

Table 8.2 reports 1991 monthly market averages for the players when identical
pricing strategies are followed that are used throughout the proof of Proposition 3. In
1991, both cooperatives operate balancing facilities, which are modeled in this game.

Table 8.2 Cooperative 1991 Monthly Averages in Million Pounds
Cooperative M Cooperative B

Class 1 Sales 90,386, 1 70 45,474,3 1 9
Balanced Milk 66,621,918 14,650,001

 

 

Employing Nash’s solution, the ﬂow of the bargaining game is as follows:
Step 1: Each player determines its threat.

Step 2: The threats are announced.

 

Step 3: Each player determines its demand as a function of all players’ threats.
Step 4: The demands are presented and if compatible, the game ends with joint action.
If the demands are incompatible, the game ends with a reversion to the status

quo.

The ﬁrst step of this game utilizes the threat-point (hB, hM) ﬁ'om Proposition 2,
illustrated using 1991 market information. Since this game is a bargaining game, the
threat-point equals the status quo (v3, vM) , and is:

(8.17) v3 = (454,743)(p — m) — 01,046,500) , and,

vM = (903,862)( p - m) — cM (666,220) ,
for one period If a cooperative implements its threat, m = p , and no over-order
premiums are generated, (8.17) simpliﬁes to:

(8.18) v8 = -cB(l46,500), and,

vM = —cM(666,220).

 

120

Although (8.18) deﬁnes the status quo, it does not complete Step 1 by deﬁning the
threats, t M and t3 , where a cooperative sets price equal to the federal order minimum.
Unfortunately, threats are not formally deﬁned without the probability of their
implementation. (In practice it is difﬁcult to interpret the threats of an opponent because
the probability of implementation depends on several factors such as cooperative debt,
patience, discount rate, preferences of c00perative management, as well as other factors.)

After the threats are presented, each player’s demand is determined as a function
ofall players’ threats. In this game the demands are dM (tM HE) and dB(tB ltM). In Step
4, the cooperatives present their demands. If the demands are compatible, the game ends
and the demands dictate the side-payments of cooperatives to distribute the PEC rents. If
the demands are incompatible, the game ends with both cooperatives carrying out their
threat, and each cooperative receiving the short-run status quo, the threat-point.

According to Nash, one rational solution exists in every two-player cooperative
bargaining game. This solution is the point within the core that maximizes the
differences of the players’ demands and their status quo. This solution is found by:

(8.19) max(dB - vBXdM —v..).

such that d3 + dM S (1,358,605) p .
The constraint requires that the sum of the demands be less than or equal to the total
premiums generated, which is the product of the Class I premium, p , and the average
1,358,605 cwt per month of Class I milk sold by cooperatives to the market. Since the
status quo is known, (8.19) may be expanded to:

(8.20) max((dB — (—146,500cB)X(dM — (—666,220cM)),

such that d, + dM 3 (1,358,605) p .

121

 

 

From this cooperative bargaining game solution, Michigan cooperatives collect and pool
Class I premiums by maximizing (8.20), changing quantities as the rmrket and season
dictates.

Using this general framework, a Michigan cooperative coalition to collect over-
order premiums is always rational. However, as reviewed in Chapter 5, for six months in
1992, Michigan cooperatives did not collect and pool Class I premiums. This game
theory model suggests the demands of the cooperatives were not compatible during these
six months. As a result, both cooperatives implemented their threats resulting in a price
war.

One potential reason for this demand incompatibility is a mistaken estimate by
Cooperative M of the probability of Cooperative B carrying out its threat. That is, in

1992 Cooperative M may not have believed Cooperative B would exit the PEC. If true,

disbelief might have led to an erroneous estimate of IR , which resulted in an irrational

demand, (1M . Additionally, over time the probability of Cooperative B’s threat might

have also changed. Regardless of the reason, these incompatible demands would have
resulted in Cooperative B exiting the PEC, implementing its threat, and starting a price
war.

In every period PEC members present their demands for the following month.
Once a break down in bargaining occurs, this process continues until the demands
become compatible. The demands of Cooperative B and Cooperative M were compatible
through May 1992. However, when bargaining for June rents, this model suggests
Cooperative M underestimated the probability of Cooperative B implementing its threat.

From June through December, the cooperatives’ demands were incompatible resulting in

122

 

a market void of over-order premiums. In December, the demands became compatible

and the PEC re-formed. The change in Cooperative B’s threat originally causing the

price war ultimately resulted in a new PEC agreement and a different distribution of PEC

rents.

Proposition 4: Without a credible threat to profits, Firm P free-rida on PE C
bargaining.

Until 1992, Michigan proprietary ﬁrms who contracted milk from independent
farmers for primarily Class I use remained outside the PEC. Because this milk was
outside the pool, so too were the potential premiums this milk could generate. In 1992,
the Michigan legislature neared approval for a state milk marketing order that would
require all Class I premiums be pooled and distributed among all Michigan dairy farmers.
The state order was not implemented because proprietary ﬁrms ‘joined’ the PEC in
December 1992. To explain proprietary ﬁrm conduct, this proposition examines the role
of the state milk marketing order on proprietary ﬁrms’ payoffs. For clarity and to provide
more detail, this proposition is presented in two parts. The ﬁrst considers proprietary
ﬁrm conduct absent the threat of the state milk marketing order. The second includes the
potential order in the proprietary ﬁrms’ decisions.

Proposition 4a: Without a policy threat to payoﬂs, Firm P rationally remains
independent of the PE C.

Proof: As established above, Michigan cooperatives rationally operate the PEC in every

period. To account for proprietary ﬁrm conduct in the Michigan market, this proposition

utilizes a two-player cooperative game. In the game, the players are the representative

proprietary ﬁrm, Firm P, and the PEC cooperative coalition as deﬁned in Proposition 3.

123

 

 

One distinction about Firm P should be reiterated at this point. This research
models Firm P as a seller of raw milk. However, in reality, Firm P contracts and secures
a major portion of its raw milk supply with Michigan dairy producers who do not belong
to either Cooperative M or Cooperative B. For this model, Firm P ‘joined’ the PEC in
1992. When Firm P joins or exits the PEC in this model, it means the milk controlled by
independent Michigan farmers is included or excluded in the monthly PEC calculations.
Since Firm P represents these farmers, the premiums accrue to Firm P.

Because a ﬁrm joins a coalition if and only if the net payoff exceeds its security
level, this proof compares Firm P’s potential PEC payoff to it’s security level. The game
presented here represents market conduct prior to 1992. If Firm P remains outside the
PEC, it free-rides on the PEC’s Class I price and retains these premiums, as was the case

between 1956 and 1992. Outside the PEC, Firm P’s security level, v,, , is the product of
the PEC Class I premium, ,0 , and its monthly Class I sales, Qf . Formally, this security
level is expressed as:

(8.21) v, = (Qf )(p),
where p and Qf are expressed on a per hundredweight basis.

If Firm P joins the PEC because it has a higher Class I utilization than the

cooperatives, its security level is reduced by the value of its side-payment to PEC

members, SP , which results in a payoff to Firm P, NP , of:

(8.22) N” = (Qfxp)—S”.

The rule governing Firm P’s participation in the grand coalition, gPEC, is:

(8.23) gPEC = {PEC,P}, ifand only if N” > v,,.

124

 

 

However, in this case, NP < vP , because no credible threat exists to reduce Firm P’s

security level outside the grand coalition. Therefore the inequality in (8.23) does not
apply, and Firm P rationally remains independent of the PEC.

However, Firm P joined the PEC in December 1992. Since this decision is
irrational in the above game, the characterization of Firm P’s conduct is not complete.
Proposition 4b modiﬁes this game to explain observed Firm P conduct.

Proposition 4b: The potential state milk marketing order resulm in rational conduct of
F irm P joining the PEC.

Proof: Proposition 4a shows without a threat to Firm P’s security level, Firm P will never

rationally join the PEC. This proposition adds to the game the policy threat of the

proposed state milk marketing order which might explain Firm P rationally joining the

PEC in 1992.

In 1980, Firm P’s Class I market share was about 10 percent, or about 15 million
pounds of sales per month. By 1992, Firm P’s market share increased to approximately
20 percent, or 30 million pounds. During this period, existing PEC members,
Cooperative B and Cooperative M, lost market share, income, and membership to Firm P,
which was free-riding on the PEC over-order premium. These losses motivated an effort
to force Firm P into the PEC through a state milk marketing order which would formalize
and institutionalize the PEC to include all Michigan Class I milk. The potential effect of
Firm P’s PEC participation was signiﬁcant. For example, in January 1992, Firm P’s
participation would have increased PEC revenues by $362,700.

The proposed order mandated that Class I premiums generated in Michigan be

shared across all Michigan milk. Furthermore, an appointed committee would administer

125

 

 

the formalized PEC, including the duty of setting the Class I premium in addition to
determining how collected premiums were divided. Clearly, such legislation could
change the outcome of the cooperative game modeled in Proposition 4a.

Speciﬁcally, the state order adds a threat to the game and alters Firm P’s security
level by making the payoffs in (8.22) permanent. The PEC agreement is not binding law,

which allows periodic deviations by Firm P to achieve the payoﬂ' illustrated by (8.21).
Under a state order, Firm P’s payoff equals its modiﬁed security level via ,

expressed as:
(8.24) vi? = 2(Qfxp)—s”.

In contrast, when Firm P voluntarily joins the PEC, using (8.21) and (8.22), its

payoff over time, It}; , is:

(8.25) 7:; = 212,?" +2 v,,,

where the ﬁrst term represents the PEC payoffs in 0 periods of voluntary PEC
‘participation,’ equal to the payoffs under the state order, via . The second term

represents payoffs during i periods of deviation from the agreement.
With the potential state order, Firm P’s decision rule regarding participation in the

grand coalition, gPEC, is:
(8.26) gPEC = {PEC, P}, ifand only if a}. >v§0.
In this game, because a]. is greater than via , Firm P rationally joins the PEC.

Of course, the decision is not quite this simple for Firm P, as the state order is not

a certainty. Assume that deviation from the voluntary PEC agreement increases the

126

 

 

 

probability of a state order in following periods because of increased cooperative
lobbying efforts. To captm'e this relationship, this proof utilizes a new term describing
Firm P’s payoff over time including the probability of the state order, CD . Speciﬁcally,

this payoff is expressed as:

(8.27) ;(¢Xvi0)-Z(1-¢va),

l

 

suchthat 0.<.<DSIand%?>0,
I

where c is the total number of periods Firm P cooperates or joins the PEC, and i is the
total number of periods Firm P acts independently or deviates ﬁ'om the PEC agreement.
That is, (I) is the probability of a state order which increases with the number of periods
Firm P deviates from the PEC agreement, i. With a low probability of (D , as (D
approaches zero, the payoff nears Firm P’s security level without the policy threat, vP ,
and Firm P exits the PEC. However, with a high probability of (D , as (D approaches 1,
the payoff resembles the payoff from a state order, Firm P losses its future opportunity
for deviation, so Firm P rationally joins the PEC ﬁom (8.26).

Generally, (D is a function of the public policy environment including the
executive, legislative, and judicial branches, as well as political prowess of Michigan
cooperatives, public opinion, and many other factors. Speciﬁcally, in periods when Firm
P deviates, i, it receives immediate payoffs associated with VP , but this deviation also
results in increased lobbying by Michigan cooperatives increasing the value of (D.

During these price wars, Firm P estimates (I). As a state order forever reduces Firm P’s
payoffs in the next period by SP , and eliminates the possibility to increase proﬁts by

free-riding, when (D exceeds some critical threshold level, (D. , Firm P voluntarily

127

 

returns to the PEC. If the probability of a state order is low, or (I) < (13' , Firm P
rationally remains independent from the PEC, and follows the decision rule of
Proposition 4a. In this way, Firm P’s deviation provides both increased payoffs and
valuable information about (I).

Using the result of this proposition, recent conduct of Firm P including the 1997

price war may be explained. In 1992, a state order neared legislative approval. This

resulted in Firm P’s voluntary participation in the PBC as (D > (1).. This action is

 

expected and rational.

This game theory model also demonstrates the 1997 price war is rational as Firm
P realizes higher payoffs outside the PEC and improves information about the value of
(I). Therefore, Firm P may have exited the PEC in 1997 after ﬁve years of voluntary
participation, to achieve higher payoffs during the peak ﬂuid milk season, observe the
reaction by the cooperatives, and form a new expectation of the state order probability,
(I). In the spring of 1998, Michigan cooperatives were moving forward with a proposed
state order, (D potentially surpassed (D. and Firm P rationally returned to the PEC.

The value of (D is unique as it depends on the memory of the experience by
participants. That is, if the memory effect is long, (D decays slowly over time. However,
if the memory effect is short, (D decays rapidly over time, leading to rational and
frequent exits by Firm P from the PEC.

8.1 Summary and Conclusions of Conduct in the Michigan Fluid Milk Market

The propositions in this chapter explain the existence of over-order premiums, the
role of the PEC, the constraints on Michigan cooperatives, why coalitions form between

cooperatives, and conduct of proprietary ﬁrms in the Michigan milk market.

128

Speciﬁcally, in the long-run without formal agreements, cooperatives collect premiums
through tacit collusion rather tlmn price ﬂuid milk at the federal order minimums.
However, formal agreements might exist because of asymmetries in cooperative
bargaining power determined by short and long-run constraints. Cartel price wars might
be caused by incompatible demands between cooperatives, which reﬂect changes in
cartel members’ demands and/or threats. These price wars are rational and result in a
new formal agreement and distribution of rents. Proprietary ﬁrms participate in the PEC
only with a credible threat to their proﬁts. They rationally join the PEC to reduce the
probability of a state order as well as secure the ﬁxture opportunity to deviate from the
formal agreement to increase proﬁts. Table 8.3 summarizes these propositions and

associated conclusions about the Michigan milk market.

129

 

Table 8.3 Summary of Propositions and Conclusions

Proposition

Conclusions

 

Proposition 1: Over-over premiums are
generated by rational decisions of
cooperatives in the Michigan ﬂuid milk
market.

Without a formal agreement, Michigan
cooperatives rationally tacitly collude in the
long-run to generate over-order premiums
Balancing functions do not change this
outcome

Rational tacit collusion may explain 26
months of over-order premiums during the
1982 price war

 

Proposition 2: The potential of a formal
agreement may give bargaining power to
Cooperative M.

The potential of a formal PEC agreement
allows each cooperative to retain a threat of
setting price equal to the federal order
minimum

Short-run constraints of one cooperative
gives bargaining power to the other, as a
short-run constrained cooperative prefers the
PEC payoffs to payoffs associated with the
threat for an unknown length of time
Cooperative B’s short-run constraints gives
bargaining power to Cooperative M,
resulting in a PEC agreement which is not
long-run rational for Cooperative B

 

Proposition 3: Incompatible demands may

be a cause of the I 992 price war. While

compatible demands might have caused the

re-organization of the PE C.

Incompatible demands within the PEC result
in price wars and implementation of each
cooperative’s threat

Compatible demands restore the PEC and
changes the distribution of PEC rents

 

Proposition 4: Without a credible threat to

proﬁts, Firm P ﬂee-rides on PE C
bargaining.

130

Proprietary ﬁrms rationally free-ride on PEC
bargaining absent of a threat to proﬁts

The potential state order serves as a policy
threat resulting in Firm P’s rational
participation in the PEC; if the potential
threat does not exist Firm P does not
participate in the PEC

Voluntary PEC participation results in
rational periodic exits by Firm P to increase
proﬁts

The probability of the state order and the
memory of the experience dictates the
frequency and length of Firm P’s exits and
returns

 

Chapter 9: Conclusions

To conclude this thesis, this chapter reviews the major ﬁndings and suggests
further research. The ﬁrst section summarizes contributions to understanding ﬂuid-milk
marketing agencies-in-common operation and policy. The second section discusses
related topics and directions for further research on the existence, operation, effects, and
stability of these marketing agencies-in-common.

9.1 Contributions

This research examined the existence, operation, eﬂ‘ects, and stability of marketing
agencies-in-common that generate over-order premiums in the ﬂuid-milk market. The
literature on milk marketing, milk supply and demand, cooperatives, federal orders, and
marketing agencies-in-common were examined to explain the existence of over-order
premiums. With suﬂicient market control, cooperatives and marketing agencies-in-
common increase the price of Class I and Class II milk. Although this price increase
results in decreased Class I and Class II consumption, it increases Class HI consumption
resulting in a higher farm blend price. This higher blend price increases production at the
farm level. However, due to the inelastic demand of Class I and Class II milk, total
market revenues increase as a result of the higher Class I and Class H prices making the
marketing agency-in-common proﬁtable.

Marketing agencies-in-common exist due to a limited legal exemption from anti-
trust laws of the Sherman Act. The Capper-Volstead Act provides farmers the
opportunity to form cooperatives and marketing agencies-in-common provided these
associations do not monopolize or restrain trade as interpreted by the US. Secretary of

Agriculture. Since 1922 no cases of monopolization or restraints of trade have been found

131

 

 

by the Secretary relative to ﬂuid-milk over-order premiums. This is not surprising as for
political reasons, it is unlikely the Secretary of Agriculture would ever rule that dairy
farmers are exploiting consumers through monopolization or restraining trade.

The history and inter-workings of an operating rmrketing agency-in-common were
also examined. The PEC has formally operated in Michigan since 1966, extracting over-
order premiums and pooling these premiums between members. This thesis chronicled the
development of Michigan’s ﬁrst marketing agency-in—common, the era of regional over-
order pools, and the formation of one superpool covering the entire Lower Peninsula in
1963.

Available data focused this analysis to the PEC’s operation since 1980. Annually,
the PEC generates more than $20 million in Class I premiums for Michigan dairy farmers.
The impact of the PEC is found by determining the cost of each price war. The cost to
Michigan dairy farmers of the 6 month 1992 price war was $14 million. The cost to the
PEC of the 1997 price war, when Firm P exited the PEC, was approximately $3.5 million.
As a result of these price wars, the pooling mechanism of the PEC was modiﬁed, changing
each market participant’s net payoﬂs. Firm P received it highest payoff under the 1985
agreement, ﬁee-riding on the PEC’s over-order premiums. Cooperative M received its
highest payoff under the 1992 agreement when it received compensation for balancing and
Firm P participated in the PEC. Cooperative B received its highest payoff under the 1997
agreement as Firm P participated in the PEC and Cooperative B did not pay for market-
wide balancing.

To complete the analysis of the PEC, this research used modern industrial

organization theory, paying particular attention to the theories of cartel stability and price

132

 

 

wars. The PEC exhibits sufficient market power to affect price, demand is sufﬁciently
inelastic, and the near-perfect information provided by the federal orders limits the need
for formal cartel enforcement policies. However, PEC price wars occur because members
desire re-negotiation of the agreement and a different distribution of the rents. Therefore,
PEC price wars are bargaining price wars.

The forrml analysis utilized cooperative and non-cooperative games to model
conduct of Michigan ﬂuid milk marketing organizations. The analysis took the form of
propositions and informal proofs.

Over-order premiums exist in Michigan without the possibility of a marketing
agency-in-common as a function of its market structure. Because two cooperatives
control about 70 percent of the Class I milk, Michigan cooperatives can rationally extract
ﬂuid over-order premiums in the long-run through tacit collusion. This conclusion rests
on the assumption that industry oligopolists rationally tacitly collude over time, rather than
price competitively, as was observed in Michigan during the 1982 price war.

Building on the rational long-run existence of over-order premiums through tacit
collusion, the second proposition showed PEC bargaining power rests with the
cooperative with the greatest short-run ﬂexibility. That is, short-run constraints restrict a
cooperative’s ability to endure another cooperative’s threat of setting Class I price equal
to the federal order price. If a cooperative is short-run constrained by its cost structure,
proﬁtability, debt load, or preferences of management, members, or the board of directors,
this lack of short-run ﬂexibility may lead a cooperative to agree to a sub-optimal long-run

agreement, which is optimal in the short-run. Therefore, a cooperative with short-run

133

ﬂexibility may rationally implement its threat and gain bargaining power over other market
participants resulting in a favorable formal agreement.

The third proposition suggested PEC price wars could occur due to incompatible
demands on the part of the players. Demands for a speciﬁc distribution of the rents are
incompatible if one player mistakenly estimates the probability of another player setting
price equal to the federal order minimum price. In this situation, each PEC member
demands a certain share of total PEC premiums, as well as retains the threat of not
collecting over-order premiums for an unknown number of periods. The implementation
of this threat rationally results in an outcome where no market participant collects over-
order premiums. Over time, players’ demands and the probability of a player
implementing its threat changes based on internal ﬁrm proﬁtability, structure, and
preference factors. Therefore, price wars are rational within the PEC which ultimately
result in new agreements and rent distribution.

The ﬁrst three propositions did not account for the participation of proprietary
ﬁrms in the PEC. Because proprietary ﬁrms can free-ride on cooperative bargaining,
voluntary participation in the PEC is not expected. However, after accounting for a
credible threat to proﬁts through the potential for a state order, voluntary participation in
the PEC is rational and expected by proprietary ﬁrms. Furthermore, because the
probability of this threat decays over time due to the memory of the experience of the
participants and a time lag exists to restore this threat, it is rational for proprietary ﬁrms to
periodically exit the PEC. In so doing, proprietary ﬁrms ﬁee-ride on cooperative
bargaining increasing proﬁts for a period of time and then return to the PEC if and when

the threat of the state order in the next period exceeds a critical level. The game theory

134

 

 

model suggested this rationale might explain the departure of proprietary ﬁrms from the
PEC in 1997, which cost the PEC over $4 million in potential revenues. Under this
scenario, if Michigan cooperatives desire stability and preservation of the PEC including
proprietary ﬁrms, they must ﬁnd a way to prevent the decay of the probability of this
threat, or implement the policy threat before the next price war.

Beyond the formal propositions, the data provided some insights into cooperative
market-wide balancing and federal order Class I differentials. During the 1982 price war,
the model assumes Class I over-order premiums fell to zero when the cooperatives moved
from tacit collusion to a state of competitive pricing in the 27th month. However, the data
reveals that during this period competitive pricing and bitterness between Michigan
cooperatives, a small over-order premium or service charge, of approximately 13 cents,
was collected by cooperatives on the sale of Class I milk. The existence of this service
charge suggests that Class I differentials do not adequately compensate cooperatives for
balancing the market. It also is evidence that milk processors understand this role of
cooperatives and pay this service charge to the cooperatives.

9.2 Further Research

As with any research there exist areas that can be identiﬁed as lurving potential for
further study. The research was constrained by data, limiting this analysis to the period
after 1980. From observation and narrative accounts, the Michigan ﬂuid-milk market was
signiﬁcantly different prior to 1980. Cooperatives had a larger share of this market and
more cooperatives existed ranging in size and the nature of the organization Therefore,
the role of the PEC agreement between 1966 and 1980 was likely very different than

today. From an industrial organization view, the PEC may have been the means to

135

 

overcome the lack of market control, the high number of ﬁrms, and information required
to sustain higher prices over time. Therefore, the PEC may have been necessary to
stabilize collusion and generate uniform over-order premiums for asymmetric
cooperatives.

Additionally, this research did not assess the role of the federal milk marketing
orders in facilitating marketing agencies-in-common. Classiﬁed pricing and the
information available as a result of the federal orders certainly assist marketing agencies-
in-common Furthermore, the PEC is not the only marketing agency-in-common in the
United States. To better understand these associations, a broader study across markets
looking at the operation and stability is needed.

Finally, more empirical work is needed related to the PEC. From the view of the
entire supply-chain, the welfare effects of the PEC are yet to be determined. However, the
welfare implications must review different levels within the market, speciﬁcally from the
PEC to milk processors, and from milk processors to retail outlets. The PEC and other
marketing agencies-in-common extract the rents otherwise captured by ﬂuid-milk

processors and retailers.

136

 

 

 

REFERENCES

137

REFERENCES

Babb, EM. and DA. Bessler. Factors Affecting Over-Order Payments in Federal Milk
Marketing Orders, 1965-80. Research Bulletin 977. Purdue Agricultural
Experiment Station. June 1983.

Bacharach, M. Economics and the Theory of Games. Westview Press: Boulder, CO.
1977.

Bailey, K.W. Marketing and Pricing of Milk and Dairy Products in the United States.
Iowa State University Press: Ames, IA. 1997.

Barton, D.G. “What is a Cooperative?” in Cooperatives in Agriculture, D.W. Cobia, Ed.,
Prentice-Hall, Inc.: Englewood Cliffs, NJ. 1989.

Carlton, D.W. and J.M. Perloff. Modern Indsutrial Organization. Harper Collins: New
York, NY. 1994.

Cropp, R.A., E.V. Jesse, and W. Dobson. “Marketing Agencies-In-Common.” Dairy
Markets and Policy Issues and Options. Cornell University’s Program in Dairy
Markets and Policy, January 1993.

Gibbons, R. Game Theory for Applied Economists. Princeton University Press: Princeton,
NJ. 1992.

Green, E.J., and RH. Porter. “Noncooperative Collusion Under Imperfect Price
Information.” Econometrica 52(January l984):87-100.

Hehnberger, P., and S. Hoos. “Cooperative Enterprise and Organization Theory.” J.
Farm Econ. 44(May 1962):275-290.

Jacobsen, R., and R.Cropp. “Dairy Cooperatives and Their Role In the United States.”
Dairy Markets and Policy Issues and Options. Cornell University’s Program in
Dairy Markets and Policy, August 1995.

Juevs, LE. and A. McMillan. “Michigan Dairy Plants-Changes in Size, Numbers and
Products Produced 1950 and 1963.” Agricultural Economic Report 4. March 1965.
Michigan State University, Department of Agricultural Economics.

Lamoreaux, N.R. The Great Merger Movement in American Business, 1895-1904.
Cambridge University Press: Cambridge, UK 1985.

Levenstein, M.C. “Price Wars and the Stability of Collusion: A Study of the Pre-World
WarI Bromine Industry.” J. of Industrial Econ. 45(June 1997):]17-37.

138

Manchester, A.C. The Public Role in the Dairy Economy: Why and How Governments
Intervene in the Milk Business. Westview Press: Boulder, CO. 1983.

Manchester, A.C. and DP Blayney. The Structure of Dairy Markets: Past, Present and
Future. United States Department of Agriculture. Agricultural Economic Report
Number 757. September 1997.

Michigan Milk Messenger. Michigan Milk Producers’ Association, Novi, MI. various
issues.

Michigan Farmer. Farm Progress Companies, Inc., Carol Stream, IL. Various issues.
Nash, J.F., “T wo-Person Cooperative Games.” Econometrica. 21(1953):128-40.

Pindyck, RS. and D.L. Rubinfeld. Microeconomics. Macmillan Publishing Company:
New York, NY. 1989.

Porter, R.H. “On the Incidence and Duration of Price Wars.” J. Industrial Econ. 33(June
1985):415-26.

Porter, R.H. “A Study of Cartel Stability: the Joint Executive Committee, 1880-1886.”
Bell J. of Econ. 14(Autumn 1983):301-14.

Rotemberg, J.J., and G. Saloner. “A Supergame-Theoretic Model of Price Wars during
Booms.” Amer. Econ. Review 76(June l986):390-407.

Scherer, F .M. Industrial Market Structure and Economic Performance. Rand McNally &
Company: Chicago, IL. 1970.

Schrader, L.F. “Economic Justiﬁcation.” in Cooperatives in Agriculture, D.W. Cobia,
Ed., Prentice-Hall, Inc.: Englewood Cliffs, NJ. 1989.

Staatz, J .M. A Theoretical Perspective on the Behavior of Farmers’ Cooperatives. Ph.D.
dissertation. Michigan State University. 1984.

Stigler, G]. “A Theory of Oligopoly.” J. Political Economy 72(February 1964):44-61.

Sumner, D.A., and C.A. Wolf. “Quotas Without Supply Control: Effects of Dairy Quota
Policy in California.” Amer. J. Agr. Econ. 78(May 1996):354-66.

United States Department of Agriculture, Agriculture Marketing Service, Dairy Division.
1989. The Federal Milk Marketing Order Program, Marketing Bulletin 27.

United States Department of Agriculture, Rural Business and Cooperative Development
Service. 1995. Dairy Cooperatives. Cooperative Information Report 1 Section 16.

139

United States Department of Agriculture, National Agricultural Statistics Service.
Agricultural Prices Annual Summary. various years.

United States Department of Agriculture, National Agricultural Statistics Service.
Agricultural Statistics. various years.

United States Department of Agriculture, Agricultural Marketing Service. Federal Milk
Order Market Statistics. various years.

United States Department of Agriculture, Agricultural Marketing Service. Market
Administrator ’s Bulletin. various years.

Wohlgenant, M.K. “Demand for Farm Output in a Complete System of Demand
Functions.” Amer. J. Agr. Econ. 71(May 1989):24l-52.

Wolf, C.A., and LG. Hamm. “The Role of Cooperatives in Milk Marketing.” Selected

Paper American Agricultural Economics Association Annual Meeting. August
1998.

140

"‘tllllillllllt