1 I
35.13:.
5.2.4..
2.5:: .
:5 z

.14

as! g

.1114. 2. .1

.b .51...
11..

 

 

 

 

 

 

 

This is to certify that the
dissertation entitled

THE EFFECTS OF COMPLIANCE UNCERTAINTY ON FLEXIBLE
ENVIRONMENTAL PERFORMANCE STANDARDS: TOTAL
MAXIMUM DAILY LOADS IN MICHIGAN

presented by

Lorie Srivastava

has been accepted towards fulfillment
of the requirements for the

 

 

 

PhD. degree in Agricultural Economics
: Major Professors éignature
i3 ' .30} L373
Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

e--‘u---.-o-.‘o-I-o-o-o-o-O-C-o-n-.--——4

 

. LIBRARY
Michigan State
University

 

 

 

 

PLACE lN 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

 

[JEWEWTZQIISS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DaleDue.p65—p.15

THE EFFECTS OF COMPLIANCE UNCERTAINTY ON FLEXIBLE
ENVIRONMENTAL PERFORMANCE STANDARDS: TOTAL MAXIMUM DAILY
LOADS IN MICHIGAN
By

Lorie Srivastava

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Agricultural Economics

2004

 

ABSTRACT
THE EFFECTS OF COMPLIANCE UNCERTAINTY ON FLEXIBLE

ENVIRONMENTAL PERFORMANCE STANDARDS: TOTAL MAXIMUM DAILY
LOADS IN IvﬂCHIGAN

By

Lorie Srivastava

Interest in ﬂexible environmental performance standards for regulating nonpoint
source pollution has been increasing. This rise in interest is partially due to such
perceived beneﬁts as the provision of dynamic incentives that allow ﬁrms to be
technologically innovative, improving the cost-effectiveness of environmental
regulations. These cost savings that may arise are due to decreases in compliance costs
for the regulated ﬁrm, and are termed innovation oﬁivets by Porter and Van der Lynde in

the Porter Hypothesis.

In the case of nonpoint sources of pollution, however, transaction costs — due to
compliance uncertainty —— may be so high that they may negate any possible beneﬁts,
limiting the feasibility of ﬂexible environmental performance standards as a policy tool.
Regulators and farmers suffer from compliance uncertainty. Regulators experience
compliance uncertainty due to (a) unobservable effluent, (b) incomplete knowledge on
their behalf with respect to understanding the fate and transport of efﬂuent from nonpoint
sources and its stochastic characteristics, and (c) opportunistic behavior by farmers due to
asymmetric information leading to moral hazard. Farmers suffer from compliance

uncertainty with respect to whether their actions or effort will be deemed sufﬁcient for

compliance with ﬂexible performance standards. Even if farmers would like to comply,
liability issues may prevent them from innovating. With moral hazard, poor
understanding of the fate and transport of pollutants, and possible free riding, farmers
may face liability issues that are not due to their own efforts. In this case, they may prefer
to forgo investments and not innovate. Furthermore, they may decide to shift liability

costs to the regulator by demanding that the regulator specify design standards.

A model is developed to examine total maximum daily loads, authorized under
the Clean Water Act to control nonpoint sources of pollution. This analysis of total
maximum daily loads, which can be implemented by ﬂexible performance standards,
reveals that transaction costs due to compliance uncertainty affect the effort level of
farmers, their costs, and their beneﬁts from induced innovation. This analysis ﬁnds that
under certain conditions, as the compliance uncertainty increases, two things occur: the
amount of farmer effort expended is greater, and the cost to the farmer increases. Under
certain conditions, cost savings from induced innovations are secondary to the cost—

savings from decreased compliance uncertainty.

Once assumptions with respect to compliance uncertainty, expected penalties, and
farmer liability that are taken for granted — implicitly or explicitly — in the customary
examination of ﬂexible performance standards are relaxed, and relevant transaction costs
are included in the analysis, ﬂexible performance standards may not be more cost-

effective than design standards.

 

Capyright by
Lorie Srivastava
2004

 

To my husband James McQueen, and to Mum and Dad, for their support and love.

ACKNOWLEDGMENTS

Probably the most important lesson that I realized in graduate school is that what
one learns is as much a function of those around you as one’s own abilities and efforts. I
was truly fortunate to have been surrounded by many talented people — faculty, fellow
students, and friends — who were willing to share their insights and knowledge with me

along this journey.

I cannot adequately express my gratitude to Dr. Batie for her guidance, her
foolproof advice, and her outstanding mentoring throughout my time at MSU. Dr. Batie
graciously supported me without hesitation right from the get go until the end. I was truly
lucky in having her as my major professor, and beneﬁted greatly from her generosity and
wisdom. I cannot convey the importance of her inﬂuence on me personally and
professionally. I would like to acknowledge ﬁnancial support from the ER. Smith

Endowment Fund and the Great Lakes Protection Fund for this research.

I owe a special thank you to Rick Horan, without whom my dissertation would be
a much inferior product. I honestly do not know whether I would have ﬁnished without

his help and meticulous attention to detail (I mean it in a good way, really!).

I would like to thank my other committee members —— Patricia Norris, Laura
Cheney, Evangelyn Alocilj a, and Chris Wolf - for being patient with me and meeting my
crazy deadlines at the end. I especially want to thank Pat and Laura for treating me more
like a friend than a student. Thank you to Gary Boersen, Dr. Dave Beade, and Steve
Davis for their assistance and insightful contributions. Thank you to Barb Dart for

arranging a meeting which ultimately led to the focus groups. It was all done on a tight

vi

schedule, and I appreciate the help of the extension agents and farmers in meeting with
me despite their other commitments. My thanks also to Corey Ri sch who helped me

conduct the focus groups.

I wish to express my gratitude to faculty members who fundamentally changed
the way I think about this discipline — Jim Shaﬁ‘er, John Staatz, A1 Schmid, Dave
Schweikhardt,.Pat Norris, and Bob Myers. Thank you Dr. Crawford for kindly listening
to all my suggestions and putting ofﬁce politics into perspective. Thank you also to Ann,

Sherry, Pat, and Nancy, and indeed all the support staff with help in navigating the rules.

I will never forget the camaraderie and friendships I made, but there are certain
friends I especially wish to mention — for listening to me vent about issues big and small
and for making my time at MSU so special: Sharon Lerner, Mollie Woods, Gregg
Hadley, Augustin Arcenas, J ing-Yi Lai, Michael Mascarenhas and Kelly Grindstaff,
Denise Mainville, Brian Radke and Natalie Hamilton, Brady Deaton and Justine

Richardson, Meeta Punjabi, Amy Damon, and Eric Knepper.

Thank you to Stanley and Joanna Stark for taking me into their home and hearts
from the ﬁrst day of my arrival. They never hesitated to help, whether it was to ﬁnd an
apartment, store a canoe, put together a Halloween costume, or more often than not, to

feed a poor graduate student.

I want to thank my parents for never doubting I would ﬁnish, and only asking on
occasion when I would. They are the reason I was able to come to MSU. Again, I
dedicate the Gs and Rs to Rolie, and the Ds and ES to Divya. Last but not least, I want to
express my love and gratitude to James, who encouraged me to go where I felt was best

for me, for moving to Michigan, for his kindness, and for ensuring I had fun.

vii

 

 

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................
LIST OF FIGURES ...........................................................................................................

Chapter 1
Introduction ........................................................................................................................
1.1 Flexible Environmental Regulations .................................................................
1.2 The Effect of Transaction Costs ........................................................................
1.2.1 Transaction Costs Incurred by Regulators .................................................
1.2.2 Transaction Costs Incurred by Farmers .....................................................
1.3 Total Maximum Daily Loads .............................................................................
1.4 Research Questions ............................................................................................
1.5' Dissertation Outline ...........................................................................................

Chapter 2
The Clean Water Act and Total Maximum Daily Loads ...................................................
2.1 History of Early Federal Water Pollution Regulations ......................................
2.2 Inclusion of Total Maximum Daily Loads in the 1972 Amendments ...............
2.2.1 The 1972 Amendments and Section 303(d) ..............................................
2.3 Implementation of the 1972 Amendments ........................................................
2.3.1 Deﬁnition of Point and Nonpoint Sources .................................................
2.3.2 Slow Implementation of Section 303(d) and Declining Water Quality
2.4 Implementation of Total Maximum Daily Loads ..............................................
2.5 The Total Maximum Daily Load Process in Michigan .....................................

Chapter 3
Uncertainty and Transaction Costs ....................................................................................
3.1 Transaction Costs and Agro-Environmental Policy ..........................................
3.1.1 Sources of Transaction Costs: Bounded Rationality .................................
3.1.2 Sources of Transaction Costs: Opportunistic Behavior .............................
3.2 Types of Transaction Costs ...............................................................................
3.2.1 Transaction Costs Incurred by Regulator .........................................................
3.2.2 Transaction Costs Incurred by Farmers ............................................................
3.3 Flexibility and Compliance Certainty Trade-off ...............................................

Chapter 4

Porter Hypothesis and Innovation Offsets .........................................................................
4.] Behavior Implications ........................................................................................
4.2 A Simple Model of Innovation Offsets .............................................................
4.3 Where Will a Farm Be? .....................................................................................
4.4 Are Innovation Offsets Possible? ......................................................................

viii

Chapter 5

A Game-Theoretic Analysis of Farmers’ Choices ............................................................. f
5.1 A Game Theoretical Framework
5.1.1 The Players and the Rules of the Game ..................................................... I
5.1.2 The Expected Payoth of the Game ........................................................... t
5.2 Sequential Extensive-Form Games .................................................................... 1
Chapter 6
Environmental Standards and Compliance Uncertainty .................................................... ’.
6.1 The Structure of the Model ................................................................................ ’.
6.1.1 Phosphorus in the Lake .............................................................................. 1
6.1.2 Information Conditions .............................................................................. '.
6-2 Performance and Design Standards with Compliance Certainty ....................... ’.
6.3 An Ambient Performance Standard with Compliance Uncertainty .................. ’.
6.3.1 Special Cases of Interest ............................................................................ l
6.4 An Eﬂluent Performance Standard with Compliance Uncertainty ................... I
6.5 A Design Standard to Reduce Compliance Uncertainty .................................... S
6.6 Implications for the Porter Hypothesis ............................................................ 1(
Chapter 7
Interview Findings ........................................................................................................... 1(
7.1 Data Collection from Regulators and Dairy Farmers ...................................... 1(
7.2 Results of Interviews with Regulators ............................................................. 1(
7.2.1 Few Implementation Details .................................................................... 1(
7.2.2 Minimizing Monitoring and Enforcement Costs ..................................... 1(
7.2.3 Move Towards a Design Standard ........................................................... 1(
7.3 Results of Dairy Farmer Focus Groups ........................................................... 11
7.3.1 Desire to Reduce Uncertainty .................................................................. 1]
7.3.2 The Paradoxical Desire for Design Standards and Flexibility ................. 11
7.3.3 Equity Concerns ....................................................................................... 11
Chapter 8
Summary and Conclusion ................................................................................................ 11
8.1 Contribution to the Literature .......................................................................... 1 I
8.2 Policy implications .......................................................................................... 12
8.3 Caveat for research .......................................................................................... 12
8.4 Future research ................................................................................................. 12
APPENDIX A
Survey with Regulators .................................................................................................... 12
APPENDIX B
Materials for Farmer Focus Groups ................................................................................. 12
REFERENCES ................................................................................................................ lf

ix

LIST OF TABLES

Table 1: Selected Ex ante Transaction Costs Incurred by Regulator ................................ 35
Table 2: Selected Ex post Transaction Costs Incurred by Regulator ................................. 36
Table 3: Relative Ranking of F irm Incentives With and Without Uncertainty ................. 84
Table 4: Comparison of Firm Incentives with Uncertainty-Reduction Effect .................. 95
Table 5: Ranking of Compliance Measures ....................................................................... 98
Table 6: Perceived Criteria Necessary to Reduce Uncertainty ........................................ 112
Table 7: Ranking of Reasons Farmers Would Comply with TMDL ............................... 113

Table 8: Weighted Ranking of Reasons Farmers Would not Comply with TMDL ........ 117

LIST OF FIGURES

Figure 1: Implementation of Total Maximum Daily Loads .............................................. 25
Figure 2: Flexibility and Compliance Uncertainty Frontier .............................................. 42
Figure 3: Proﬁts and Environmental Quality Frontier ....................................................... 50
Figure 4: Induced Innovation and Innovation Offsets ....................................................... 54
Figure 5: Possible Actions and Outcomes of a TMDL with Compliance Uncertainty ..... 64
Figure 6: Effort and Innovation with Perfect Information ................................................. 76
Figure 7: Eﬁ‘ort and Innovation with a Flexible Ambient Performance Standard ............ 81
Figure 8: Effort and Innovation with a Flexible Efﬂuent Performance Standard ............. 92
Figure 9: Effort and Costs with Uncertainty-Reduction Effect ......................................... 94
Figure 10: Effort and Costs with a Design Standard ....................................................... 100

xi

CHAPTER 1

Introduction

Nonpoint sources of pollution have been identiﬁed by the Environmental
Protection Agency (EPA) to be the main cause of the remaining surface water quality
problems in the United States (EPA 1998). Furthermore, since agricultural production is
the largest nonpoint source of water pollution in the US. (EPA 1998), it is being
subjected to increasing policy attention. This policy attention has resulted in the search
for an appropriate set of regulatory tools and programs to reduce agriculture’s negative
effect on water quality. Certain types of environmental regulatory measures have been
promoted over others by most economists. For example, ﬂexible environmental
performance standards are believed to be potentially more cost-effective than the
regulatory measures that have been historically employed in the United States. Most
proponents of these measures, however, have not adequately accounted for uncertainty
and its resulting transaction costs. This study examines these issues with respect to
nonpoint sources of pollution in Michigan in the context of the Clean Water Act’s Total

Maximum Daily Loads (TMDLS) program.

1.1 Flexible Environmental Regulations

Historically, environmental management in the US has been based upon a
command-and—controll philosophy. A command-and-control regulation would be the

case where the regulations specify both “what is to be done” and “how” the goal is to be

 

1 Some authors have argued that the nomenclature surrounding environmental policies is not truly
reﬂective of environmental regulations. For instance, Russell and Powell contend that command-and-
control is a pejorative term that does not adequately describe the range of environmental policies, since few
regulations specify both what is to be achieved (i.e. command) and how it is to be achieved (i.e. controlled)
(Russell and Powell 1996). See also Russell 2001. Driesen argues that the command and control/economic
incentive dichotomy must be replaced with a more nuanced analytical approach to both types of programs
in order to create a more dynamic and effective environmental law (Driesen 1998).

achieved. An example would be the requirement of a particular air scrubber to meet an air
emission standard or a prohibition of a polluting activity. Although this approach has led
to environmental improvements over the last three decades (Adler 1999; Houck 1999),
economists have criticized the command-and-control framework for being expensive,
inﬂexible, and more importantly, for stiﬂing technological innovations. As a result,
policymakers have begun searching for new pollution prevention approaches; and they

have focused increased attention on programs based upon ﬂexible incentives.

It has been argued that under certain conditions, agro-environmental policies that
incorporate ﬂexible incentives may allow farmers to attain desired environmental goals in
a cost-effective manner (Batie and Ervin 1999; Segerson 1999a; Portney and Stavins
2000). Flexible incentives are “environmental management tools that specify objectives
but do not dictate how the environmental objective is to be achieved” (Batie and Ervin
1999, p. 56). This ﬂexibility has the potential to allow policymakers to accommodate
diverse natural resource conditions, heterogeneous production systems, and other factors
that may vary spatially and temporally. Furthermore, ﬂexible incentive-based
environmental regulations remove the onus on regulators to determine the “best”

technology that must be used by the regulated ﬁrms.

One policy instrument that incorporates this ﬂexibility is a ﬂexible environmental
performance standard. An environmental performance standard establishes the
environmental limit for potential pollutants. Typically, an environmental performance
standard is designed to achieve a policy target, where the goal is based upon either human
health criterion, or an ecosystem criterion, or a combination of both. For instance, with

water quality, an ambient performance standard speciﬁes the ambient concentration of a

pollutant, such as phosphorus, in a body of water. Deviations above the allowed

concentration result in penalties.

Flexible environmental performance standards are differentiated from other types
of environmental standards. With a design standard, particular technologies must be
adopted by the polluter are prescribed;2 an example in agriculture would be the required
adoption of conservation tillage. In contrast, a technology based performance standard
refers to the amount of pollutant allowed in an efﬂuent, but it is based upon the best
outcome of a given technology. Eﬂluent performance standards are source speciﬁc and
refer to the allowable level of a pollutant that a polluter can emit. Finally, an ambient

performance standard refers to outcomes such as ambient water quality.

With a ﬂexible environmental performance standard, no design standards or
speciﬁc technologies are prescribed. Thus, ﬂexible ambient and efﬂuent performance
standards are both examples of a ﬂexible environmental performance standard. A ﬂexible
environmental performance standard allows farmers to implement the technology or

production method that is most cost-effective for their situation.

The two most signiﬁcant potential advantages of ﬂexible environmental
performance standards over the traditional command-and-control approach are: (1) they
create dynamic incentives for technological innovation and its diffusion, and (2) they
may be more cost-effective (Stavins 2000). In some cases, this dynamic incentive for
technological innovation for nonpoint pollution may even allow farmers to increase their

proﬁts. The concurrent outcomes of environmental improvements and a reduction, or

 

2 Design standards are also referred to as “technology speciﬁcations” by others and occasionally the term
design standard refers to a performance standard (e.g. Russell 2001). Throughout this study, however, a
design standard refers to a prescribed technology by the regulator.

offset, in compliance costs through the adoption of new technology is deﬁned as an
innovation oﬁfset. If farmers are actually able to generate innovation offsets while
meeting an environmental performance standard, then their costs to comply with the
environmental regulation may decrease, making the ﬂexible environmental performance
standard — either ambient or efﬂuent — potentially more attractive to farmers than

environmental regulations based upon design standards.

The potential for innovation offsets is an important implication of the “Porter
Hypothesis” (Porter and Van der Linde 1995). According to the Porter Hypothesis,
“policies mandating strict environmental compliance have [the] potential to make
American ﬁrms and industries more competitive” by encouraging technological
innovations and change (Thurow and Holt 1997 p. 20). By inducing ﬁrms to innovate, the
Porter Hypothesis contends that environmental regulations can provide necessary
dynamic incentives for technological innovations. Researchers have debated as to
whether the Porter Hypothesis does in fact hold (e. g. Gardiner and Portney 1994; Palmer,
Oates, and Portney 1995; Jaffe and Palmer 1997). These studies, however, have

examined the manufacturing sector; they have not focused on agriculture.

One of the few journal articles that examines innovation offsets in the context of
animal agriculture is by Thurow and Holt (1997). Dairy farmers were required to adopt
certain technologies to meet water quality regulations in the Okeechobee, Florida. Thus,
ﬂexibility was not embodied in the environmental legislation. But some farmers also

adopted non-required production-enhancing technologies.3 Thurow and Holt, however,

 

3 Boggess, Johns, and Meline (1997) state that compliance costs $1.14 per hundred weight of milk were
offset by a 13 per cent increase in milk production when farmers complied with the Dairy Rule in Florida.
But they did not examine induced innovation or innovation offsets speciﬁcally in this article.

found that, although these investments increased revenue, they were not associated with
phosphorus runoff reduction; consequently these innovations did not result in innovation
offsets, as deﬁned above. This study departs from Thurow and Holt’s examination in that
innovation offsets in the context of ﬂexible environmental performance standards in
agriculture are investigated, with a focus on the effects of compliance uncertainty and
resulting transaction costs. The basic thesis underlying this focus is that the incentive to
innovate as a result of environmental performance standards may be undermined when

signiﬁcant transaction costs exist.

1.2 The Effect of Transaction Costs

A transaction cost can be deﬁned to include “the costs of gathering and
processing the information needed to carry out a transaction, of reaching decisions, of
negotiating contracts, and of policing and enforcing those contracts” (Williamson 1981).
In other words, transaction costs are the costs of coordinating and motivating an

economic system (Milgrom and Roberts 1992).

Whether ﬂexible performance standards can achieve environmental outcomes at a
lower cost compared to design standards depends upon transaction costs (Batie and Ervin
1999; Norris and Thurow 1999). But what type of transaction costs are relevant? With
ﬂexible environmental performance standards, transaction costs associated with
coordination and motivation — or incentives — are of signiﬁcance; speciﬁcally, these
transaction costs are due to informational incompleteness and informational asymmetries
(Milgrom and Roberts 1992). Transaction costs arise because of a lack of sufﬁcient
information — the pervasive uncertainty in the world prevents ﬁrms and individuals from

making perfectly informed decisions. These informational problems lead to compliance

 

uncertainty which complicates the implementation of ﬂexible environmental performance

standards.

1.2.1 Transaction Costs Incurred by Regulators

With nonpoint source pollution, compliance uncertainty stems from a variety of
factors attributable to imperfect information and the stochastic nature of nonpoint source
pollution. For regulators, transaction costs such as information, monitoring, and
enforcement costs can be signiﬁcant with ﬂexible environmental performance standards.
Regulators have imperfect information with respect to assessing the efforts undertaken by
farmers. This informational constraint, or uncertainty, can be due to such factors as
technological or personnel limitations. Additionally, with nonpoint sources, the effects of
discharges are not often observed immediately; these effects can be spatially and
temporally removed from the pollutant source. This diffuse and stochastic characteristic
of nonpoint source pollution is the cause of an additional informational problem — namely
the limited understanding of the transport and fate of pollutants from nonpoint sources.
This informational constraint seriously constrains the regulator’s ability to determine the
farmer responsible for nonpoint pollution and about his degree of responsibility (Shortle,

Horan, Abler 1998).4

This high degree of compliance uncertainty may lead to moral hazard. Moral
hazard arises when the regulator can not observe the actions of individual farmers and
thus can not determine the efﬂuent from each farm, nor each farmer’s pollution

abatement efforts (Xepapadeas 1991; Tirole 1995). Rather, the regulator can only observe

‘

4 The author recognizes that there are many women farmers, but the masculine pronoun will be used for
S i mplicity.

 

 

 

 

the outcome of all farmers’ actions in the ambient pollution level of a body of water, for
instance. Essentially, unobservable efﬂuent from farm operations mean that farmers’
performance in the public interest can not be directly observed (Shortle, Horan, Abler
1998). Consequently, each farmer can increase his proﬁts by choosing a lower level of
abatement effort, since his actions can not be monitored effectively by the regulator

(Xepapadeas 1991).

Given these transaction costs, it can be challenging for regulators to even
determine what “being in compliance” means from an enforcement perspective. When
ﬂexible performance standards are applied to agricultural nonpoint source pollution,
regulators do not specify technologies to be used by regulated farms; consequently the
regulator does not check the farm operation to see whether a speciﬁc technology is being
used. The regulator may only be able to check indirectly whether a ﬁrm is in compliance.
For instance, the regulator may sample the impaired water body to measure the ambient
pollutant level. If the ambient level exceeds the speciﬁed standard, the regulator only
knows that the aggregate efﬂuent level must be reduced; this observation does not help
the regulator answer questions like: “Which farms, if any, are responsible for this

outcome?” and “How much must each farm cutback?”

Although the regulator could use a computer model to estimate efﬂuent from
farms, this method entails costs for the regulator. For example, the soﬁware has to be
developed and maintained by technically trained staff, and data must be gathered.
Scientiﬁc information, which must exist on the linkages between farm practices and

ambient water quality, may be absent. Consequently, it may be difﬁcult for the regulator

to deﬁne compliance at an operational enforcement level. In addition, the use of a

 

 

 

computer model to estimate efﬂuent and determine ﬁnes may not be defendable in court.
Compliance uncertainty thus poses a challenge for regulators in managing nonpoint

sources of pollution with a ﬂexible performance standard.

In contrast, design standards tend to be easier to administer, thereby reducing
monitoring costs for the regulator. Once the design standard is speciﬁed, the regulator
simply has to observe whether each farm is in fact employing the required technology. If
a farm is not, it is in violation of the environmental regulation; if it does use the
technology, then the farm is in compliance. The actual ambient pollutant level is of
secondary importance with regulations that employ design standards. Administration and
political difﬁculties may arise, however, if the water quality fails to improve, and the

regulator, following an adaptive management path, “readjusts” the required technology.

This lack of focus on environmental outcomes with design standards is not a
surprising result; when measuring performance for individual agents is costly, then
decision-makers create institutions that reduce the importance of accurate performance
measurements. That is, procedures-based approaches, such as design standards, are
desirable when uncertainty and complexity make it difﬁcult to predict performance
(Milgrom and Roberts 1992). This observation provides an important explanation as to
why, in the context of nonpoint source pollution, design standards have been the

preferred policy instrument in the US.

On the other hand, ﬂexible environmental performance standards may reduce
some informational costs for regulators relative to design standards. With ﬂexible
performance standards, regulators do not have to bear search costs nor undertake research

efforts of their own to determine the best technology that farmers must use to reduce

 

nonpoint source pollution. With ﬂexible performance standards, farmers now have to
determine the best technology or production method to use for their own situation.
Consequently, ﬂexible performance standards allow the regulator to shift these
transaction costs onto farmers. This reduction in informational costs may become
signiﬁcant in the long term as environmental constraints and technology change and

become more complex.

1.2.2 Transaction Costs Incurred by Farmers

There is an advantage from ﬂexible environmental performance standards for
farmers; farmers can choose the technology that is best suited to their operation and
surrounding ecosystem requirements. A farmer does not have to implement a technology
that may be too expensive or inappropriate for his circumstances, and may be able to

beneﬁt from innovation offsets from his technological choice.

Nevertheless, ﬂexible environmental performance standards also result in
transaction costs for farmers. Informational costs dominate. In order to comply with
ﬂexible performance standards, farmers must possess knowledge about the nature and
extent of efﬂuent from their farm at their source, as well as the relative cost-effectiveness
of different pollution prevention management practices. In fact, many resource
management decisions that can have a signiﬁcant effect on the environment are too costly
to monitor or verify (Shortle, Horan, Abler 1998), especially for farmers who are not
trained in determining and assessing technical scientiﬁc information. This information
can be made available to the farmer — for instance, through a computer model. But, at the
least, the farmer must learn how to interpret the results of the model, and more

importantly, he must trust the results of the model.

 

 

Stochastic weather events can compound the uncertainty surrounding links
between management practices and environmental outcomes. In fact, it has long been
recognized that nonpoint source pollution is inherently stochastic (Shortle and Dunn
1986). The diffuse nature of nonpoint source pollution compounded by stochastic
weather events also makes it difﬁcult for farmers to determine whether they are in
compliance with the ﬂexible performance standard. Consequently, the farmer may
unintentionally cause environmental degradation as a result of his management practices
since he can not link his own management practices to any deviation from the

environmental performance standard.

Thus, in practice, it is difﬁcult for farmers to ascertain whether they are in
compliance with a ﬂexible environmental performance standard due to the nature of
nonpoint source pollution. This difﬁculty can be contrasted with compliance with
environmental regulations based upon design standards. In the case of design standards,
the farmer is told what pollution prevention technology to use, so he clearly knows
whether he is in compliance. If he does not adopt the speciﬁed technology, then he is not
in compliance, otherwise he is in compliance. Thus, certain informational transaction
costs that the farmer must bear with ﬂexible performance standards are eliminated with

environmental regulations based upon design standards.

It would appear that economic theory suggests that the feasibility of ﬂexible
environmental performance standards to control or minimize nonpoint source pollution is
ambiguous. On the one hand, such standards provide incentives for technological
innovation, and they may reduce compliance costs through ensuing innovation offsets; on

the other hand, transaction costs due to compliance uncertainty may be so high with

10

 

 

 

respect to nonpoint sources, that they may negate any possible beneﬁts, rendering ﬂexible
environmental performance standards infeasible as a policy tool to control nonpoint

sources of pollution. The paucity of empirical experiences has limited the resolution of

this ambiguity.

Clearly, compliance uncertainty creates transaction costs for regulators and
farmers, since both agents have difﬁculty linking farmers’ actions with compliance to the
environmental performance standard. This compliance uncertainty and the resulting
moral hazard may prevent farmers from complying with the ﬂexible performance

standard, thereby reducing the likelihood that water quality goals will be met.

1.3 Total Maximum Daily Loads

The United States Environmental Protection Agency is now attempting to control
nonpoint sources of pollution throughout the country under provisions of the Clean Water
Act by way of a process called total maximum daily loads (TMDLs). TMDLs can be
considered to be ﬂexible environmental performances standards for surface water.5
Consequently, Michigan, along with other states with delegated authority, are beginning
to implement TMDLs to meet state water quality goals and comply with the Clean Water
Act. In fact, given the rise in the use of ﬂexible performance standards through the Clean
Water Act, it would appear that federal policymakers are undaunted by the transaction
costs associated with them, or feel that they are low enough so as to not be a signiﬁcant
impediment, or hope that transaction costs will be quickly lowered in time through

advances in information and monitoring technology. This study will examine whether

 

5 TMDLs have been authorized since 1972 in the Clean Water Act, but they have only recently been
implemented. This is explained in further detail in Chapter 2.

ll

ﬂexible performance standards are effective in controlling nonpoint sources of pollution

in the presence of transaction costs related to uncertainty.

Nutrients from nonpoint sources are major pollutants that impair lakes and
reservoirs throughout the nation, and agriculture is the leading source of pollution in
assessed surface waters (EPA 1998). In fact, agriculture contributes to 59 percent of
reported water quality problems in impaired rivers and streams (EPA 1998). The
pollutant of interest in this study with respect to total maximum daily loads is
phosphorus. Most phosphorus from agricultural lands is transported with sediment since
it binds to soil (Parker 2000). Recent research, however, has indicated that soluble
phosphorus runoff can result if the amount of phosphorus bound to soils is very high
(Sims 1998). These two facts compound the effect of stochastic events such as weather
and the transport and fate of phosphorus, and thus the control of phosphorus from

nonpoint sources.

Phosphorous is of interest because it becomes a pollutant when it enters surface
waters in substantial amounts; it contributes to the excessive growth of algae and other
aquatic vegetation, leading to the accelerated eutrophication of lake habitats and
ecosystems. Eutrophication is the process by which a body of water becomes rich in
dissolved nutrients, and hence deﬁcient in dissolved oxygen. This resultant oxygen
depletion can lead to ﬁsh kills and other water quality problems. In fact, phosphorous is
most often the limiting nutrient in freshwater aquatic systems (NRC 2001). Water bodies
impaired due to excess phosphorus are costly and difﬁcult to restore and take many years

to recover (Lemunyon and Daniel 1998). Thus, excessive phosphorous can damage

12

intricate habitat interrelationships, thereby degrading the complex Great Lakes

ecosystem.

Little empirical research has been conducted to resolve this lack of knowledge
(Schmitz et al. 1995; Thurow and Holt 1997). Although data shortcomings were
highlighted by both the General Accounting Ofﬁce and the National Research Council in
their assessments of TMDLs, the effect of compliance uncertainty on the successful
implementation of TMDLs and the issue of innovation offsets were not speciﬁcally
examined (GAO 2000, NRC 2001). Consequently, this study contributes to this nascent
literature by examining whether ﬂexible performance-based agro-environmental policies
will effectively address nonpoint sources of pollution in meeting water quality goals in

Michigan.

1.4 Research Questions

Flexible environmental performance standards are potentially more cost-effective
than design standards and thus tend to be preferred by economists. They have the
potential to be more cost effective because, as noted by the Porter Hypothesis, the
ﬂexibility makes possible induced innovation and innovation offsets. Innovation offsets
may reduce compliance costs for regulated ﬁrms, and they potentially reduce search and
informational costs for the regulator, since the regulator no longer has to determine and

specify a design standard for regulated ﬁrms.

Despite the increasing regulatory interest in ﬂexible environmental performance
standards, there is a dearth of knowledge as to how cost-effective they will be in
controlling nonpoint sources of pollution when compliance uncertainty is taken into

account. This study examines the effect of uncertainty and its resulting transaction costs

13

on the cost-effectiveness of ﬂexible environmental performance standards to curb
nonpoint source pollution. This examination will be guided by an ex ante analysis of the
implementation of total maximum daily loads by Michigan regulators to control
phosphorus. This study is an ex ante study because Michigan has yet to implement a
TMDL process. This analysis can provide insights into the regulatory process that
attempts to reduce nonpoint source pollution from agriculture in order to meet water

quality goals in Michigan.
As such, the two key research questions that will be examined are:

1. How might compliance uncertainty affect the effort level of Michigan
livestock farmers, their costs, and their benefits from induced

innovation?

2. Can ﬂexible environmental performance standards ensure induced
innovation in the context of compliance uncertainty, and what are the

implications for the Porter Hypothesis?

These research questions will be analyzed in terms of total maximum daily loads,
which if implemented as ﬂexible environmental performance standards, are the only
example of ﬂexible performance standards that currently exist for agriculture in

Michigan.

The agricultural sub-sector to be examined is the dairy sub-sector. This study
focuses on the Michigan dairy industry because it’s the highest ranked agriculture sub-
sector in terms of market value of agriculture sales, contributing to 18 per cent of all

agricultural sales in the state. Furthermore, the dairy sub-sector will be affected by the

14

 

 

implementation of total maximum daily loads as it may be a signiﬁcant source of
phosphorus (Census of Agriculture 1997).6 With approximately 300,000 dairy cattle,
Michigan ranks eighth in the country in terms of numbers of dairy cattle. Once
assumptions with respect to compliance uncertainty that are taken for granted — implicitly
or explicitly — in the customary analysis of ﬂexible performance standards are relaxed,
and the transaction costs are included in the analysis, ﬂexible performance standards may

not be more cost-effective than design standards.

1.5 Dissertation Outline

This dissertation proceeds by providing a foundation of the effects of compliance
uncertainty on the implementation of ﬂexible environmental performance standards by
examining key components in each of Chapters 2 through Chapters 5. Speciﬁcally,
Chapter 2 provides an historical overview of the institutional context, that is, of the Clean
Water Act and TMDLs. The effect of uncertainty and its resulting transaction costs on
both regulators and farmers with respect to ﬂexible performance standards is then
analyzed in detail in Chapter 3. Chapter 4 examines the Porter Hypothesis component,
speciﬁcally, the innovation offsets that may be induced by ﬂexible performance
standards. These components are weaved together into a game theoretical framework in
Chapter 5 to stylistically depict the regulator and the possible actions of a farmer who
must contend with a ﬂexible environmental performance standard in the face of
compliance uncertainty. Chapter 6 builds upon this framework to develop a conceptual
model to examine the differing levels of cost-savings due to farmer innovation that

potentially result from various levels of compliance uncertainty with a ﬂexible

 

6 Information on the actual amount of phosphorus from dairy farms or other livestock farms in Michigan is
limited.

15

performance standard. Speciﬁcally, this model is used to investigate the implementation
of a TMDL to control phosphorus from dairy nonpoint sources in the presence of
compliance uncertainty. Using data collected through interviews with state regulators and
focus groups with Michigan dairy farmers, empirical corroboration of the results from
Chapter 6 is presented in Chapter 7. Finally, Chapter 8 summarizes the major ﬁndings of

this study. Policy and research implications of this analysis are discussed as well.

 

CHAPTER 2
The Clean Water Act and Total Maximum Daily Loads

If total maximum daily loads —— implemented as ﬂexible performance standards —
can cost-effectively address the degradation of water quality in the Great Lakes
ecosystem in Michigan from the livestock agricultural sub-sector, then environmental
quality policymakers can be encouraged that they are following the right course. But if
TMDLs do hold such a promise, why were they not implemented before? Why are they
just being implemented now? How are they being implemented? To answer these

questions, a brief overview of federal U.S. water regulations is necessary.

2.1 History of Early Federal Water Pollution Regulations

Federal involvement in controlling water pollution began in 1899 with the Rivers
and Harbors Act — commonly known as the Refuse Act.7 This act, however, did not
address water pollution as thought of today; rather, it prohibited the discharge of refuse of
any kind into the navigable waters of the United States. Congress enacted the ﬁrst federal
legislation to explicitly contend with water pollution in 1948 — the Federal Water
Pollution Control Act (FWPCA). The federal government, however, had no authority to
establish water quality standards, limit discharges, nor engage in any enforcement.

Indeed, the Act speciﬁcally recognized states’ primary authority over water pollution.

The F WPCA was amended ﬁve times between 1948 and 1972. Water quality
standards were ﬁrst included in federal law in amendments in 1965.8 With these

amendments, federally approved water quality standards were required for interstate

 

7 This section is adapted from Freeman 2000, and Battle and Lipeles 1998.
8 The Water Quality Act of 1965.

17

waters.9 Nevertheless, by 1970 it was apparent that the F WPCA was ineffective in
improving the waters of the United States. A dramatic symbol of the lack of success of
the FWPCA occurred on June 22, 1969, when the heavily polluted Cuyahoga River in

Cleveland burst into ﬂames.

The revisions that Congress made to the F WPCA in 1972 were signiﬁcant. The
Senate committee that reported the 1972 amendments found that “there had been an
almost total lack of enforcement of the FWPCA and that new measures were required to
curtail the use of rivers, lakes, and streams as waste treatment systems” (Battle and

Lipeles 1998, p. 13).

2.2 Inclusion of Total Maximum Daily Loads in the 1972 Amendments

In 1972, due to reports of deteriorating water quality from around the US, a new
political view began to prevail — water pollution was a national problem that required.
federal involvement (Houck 1999). In drafting their versions of the Act, however, the

Senate and House committees took different approaches.

As Houck (1999) recounts, a number of Senators felt that because water quality
standards had completely failed as a policy tool since ﬁrst enacted in federal law in 1965,
they should be abandoned, and only technology-based standards should be in the Act.
Thus, the Senate promoted technology-based standards in its version of the new federal
water act (S. 2770). States and industry representatives, however, did not want to
implement technology standards. As the House was historically more inﬂuenced by state

and industry views, it passed H.R. 11896, retaining the emphasis on state primacy and

 

9 This surrrrnary draws heavily from Houck (1999). For a thorough legal history of the Act, see Houck.

l8

 

 

water quality standards. In order to make its water quality standards program more

defensible, however, it added Section 303(d), among other provisions.

2.2.1 The 1972 Amendments and Section 303(d)

The role of state water quality standards — pushed by the House — versus
technology standards — pushed by the Senate — was at the heart of the debates of the 1972
Amendments. The Senate did not yield on the technology approach, but, in the end, it did
compromise. House Section 303, with its water quality criteria and standards, total
maximum daily loads for polluted waters, and implementation plans was adopted, but
only as a backup if technology standards proved to be insufﬁcient in meeting water
quality goals. That is, Section 303(d) added a prescription for using water quality
standards to improve those waters that remained polluted after the application of

technology-based requirements, through the use of total maximum daily loads.

This section was a conscious response by the House to address the perceived
failings of the FWPCA, as well as recognition of the strong criticism that state
performance and water quality standards regulation were receiving from the Senate and a
strong minority of House members (Houck 1999). Additionally, there was support for
retaining water quality provisions by some environmental groups who did not want to

rely solely on technology-based standards.

Brieﬂy, Section 303 ((1) requires the states to undertake three steps: 1) identify
waters that are, and will, remain polluted after the application of technology standards, 2)
prioritize these waters, accounting for the severity of their pollution, and 3) establish total
maximum daily loads for these waters at levels necessary to meet the applicable state

water quality standards. The total maximum daily loads must account for seasonal

l9

variations and allow a margin of safety to reﬂect the lack of certainty about discharges

and water quality (Houck 1999).

There were essentially two objectives behind Section 303(d): 1) to involve both
the states and the EPA, and 2) to specify the steps required to improve degraded waters.
Thus, Section 303(d) provided both a mandate and a framework for using water quality
standards to improve polluted waters, a departure from past FWPCA amendments

(Houck 1999).

Despite its compromise, the Senate did not take Section 303 seriously; indeed,
Senator Muskie, principal author of the Senate bill, told the EPA administrator to “assign
secondary priority” to Section 303 when it came to allocation in the years ahead (Houck
1999, p. 24). Thus, the EPA and the states essentially ignored Section 303 for decades,

and focused on technology-based standards.

2.3 Implementation of the 1972 Amendments

Following the passage of the F WPCA Amendments of 1972, the EPA was fully
occupied in promulgating technology standards under the CWA and defending them in
court. The Agency had little inclination to implement Section 303 (d) before technology
requirements were in place (Houck 1999, p. 49). In fact, the EPA focused on technology

standards only for point sources, and virtually ignored nonpoint sources (Houck 1999).

2.3.1 Definition of Point and Nonpoint Sources

The Clean Water Act deﬁnes a point source to be “any discernible, conﬁned and
discrete conveyance, such as a pipe, ditch, channel, tunnel, conduit, discrete ﬁssure, or

container. It also includes vessels or other ﬂoating craft from which pollutants are or may

20

be discharged. By law, the term point source also includes concentrated animal feeding
operations...” (EPA 2002). Concentrated animal feeding operations (CAFOs) are
currently deﬁned as 1,000 animal units, which translates into 700 milking cows. Thus,
under the federal Clean Water Act, all dairy operations with 700 milking cows or more

are considered a point source under the Clean Water Act, and are regulated entities.

The EPA has deﬁned nonpoint source pollution as pollution that “is caused by
diffuse sources that are not regulated as point sources and normally is associated with
agricultural, silvicultural and urban runoff, runoff from construction activities, etc. Such
pollution results from human-made or human-induced alteration of the chemical,
physical, biological, and radiological integrity of water. In practical terms, nonpoint
source pollution does not result from a discharge at a speciﬁc, single location (such as a
single pipe) but generally results from land runoff, precipitation, atmospheric deposition,
or percolation” (EPA 1987). The term nonpoint has commonly come to be used when

referring to pollution from agricultural sources.

In 1977, Congress included an exemption for farms in the Clean Water Act, and
redeﬁned point source to exclude return ﬂows from irrigated agriculture. In 1987,
Congress exempted all agricultural storrnwater discharges. Due to these exemptions, from
the standpoint of the Clean Water Act, farms have been free to discharge soils, animal
wastes, fertilizers and other pollutants into surface waters under the Clean Water Act

(Ruhl 2000); however, other legislation, such as state rules, may apply to such efﬂuent.10

 

10 In Michigan, the Natural Resource and Environmental Protection Act (1994) prohibits such discharges,
but investigations of violations are complaint-driven. Critics claim that there are many unregulated releases
of livestock-related pollutants into Michigan waters (Sierra Club, 2000)

21

Thus, with the exception of concentrated animal feeding operations, few restrictions on

agriculture speciﬁed by the Clean Water Act were enforced.11

2.3.2 Slow Implementation of Section 303(d) and Declining Water Quality

The Act speciﬁed that Section 303(d) could be implemented only after the EPA
formally identiﬁed pollutants appropriate for water quality analysis and TMDLs. Once
these pollutants were identiﬁed, states had 180 days to submit their lists of degraded
waters, priorities for cleanup, and TMDLs. Yet, the EPA did not identify the pollutants.
Section 303(d) was ignored for two reasons: 1) assessing compliance was difﬁcult, and 2)
ignoring it seemed possible (Houck 1997). Lawsuits were ﬁled in the late 19703,
however, challenging the absence of TMDLs for parts of the Colorado River and South

Dakota, but they failed on the grounds that the EPA had not identiﬁed the pollutants.12

A court order in 1978, however, required the EPA to publish a ﬁnal identiﬁcation
of TMDL pollutants. In the notice of its required regulations, the EPA explained that it
did not consider this list to be a high priority, since it felt that many of the practical
results of TMDLs were being accomplished through basin planning (Houck 1999).The
EPA told states that nonpoint source contributions did not need to be considered in
setting priorities since the “relative signiﬁcance of point and nonpoint sources would not
be determined until TMDLs were developed.” (Houck 1999, p. 51). Thus, due to the low

priority given to TMDLs by the EPA, along with its preoccupation with technology

 

1 1 This situation continues to largely be true today.

12 The Environmental Defense Fund ﬁled a lawsuit, but the EPA did not specify salinity as a pollutant until
December 1978, allowing states until June 1979 to submit TMDL calculations. Similarly, the case in South
Dakota also failed since the EPA did not identify pollutants.

22

 

 

standards and controlling point sources, states did little or nothing to address nonpoint

SOLII'CCS.

But by the later 19805, a consensus was growing among states, Congress,
environmentalists, and industry over the need for improvements to the CWA; a primary
focus was the return to water quality standards for a chronic and unsolved problem —

namely nonpoint source pollution.13

At the same time, was becoming clear that technological controls on nonpoint
sources of pollution was insufﬁcient for waters affected by multiple dischargers (some
regulated, others not) and multiple pollutants. For instance, a program in 1972 to address
nonpoint source pollution via state and watershed planning through the CWA only

produced studies, but no measurable improvements in water quality (Houck 1999).

2.4 Implementation of Total Maximum Daily Loads

Throughout the 19808, state agencies and the regulated community continued to
insist that water quality standards should be used to control water pollution; this
insistence was unnecessary since they had the authority and mandate to do so under
Section 303(d). But Section 303(d) was largely ignored for the ﬁrst twenty years — since
the 1972 amendments to the Clean Water Act — due to inaction by the states and the lack
of enforcement pressure by the EPA (Battle and Lipeles 1998; Stephenson, Shabman,
Geyer 1999; Ruhl 2000). Most states focused on bringing point sources of pollution into
compliance (NRC 2001). States did not undertake any of the steps speciﬁed under

Section 303(d); they did not: 1) submit inventories of polluted waters, 2) they did not

 

13 Toxic water pollution was the other area problematic area of concern.

23

prioritize these waters for cleanup, and 3) they did not promulgate TMDLs. In the 19803,
citizens’ groups’ — or third-party groups’ — lawsuits accelerated, accusing the EPA of

failing to implement and enforce the total maximum daily load process (NRC 2001).

Yet the EPA did not readily comply. When the EPA was initially sued in the early
19808, it argued that Congressional intent was to rely on the states to set the TMDL
machinery in motion because, under Section 303(d), there is no explicit requirement for
the EPA to act in the absence of state inaction. That is, the EPA argued that their
intervention is called for only in response to inadequate state performance, but not in

response to no state performance. (Houck 1999).

Thus, the EPA and states effectively continued to ignore TMDLs until they were
faced with a “virtual avalanche of litigation in the mid-19903” (Battle and Lipeles
1998).14 This wave of lawsuits established that ignoring Section 303 was no longer
possible, and that continued state inaction constituted action, requiring EPA to respond.
The issues in these cases have tracked the literal requirements of the statute, ﬁrst raising
the failure to list state waters, then the inadequacy of these lists, and then the failure to
prepare TMDLs.15 To date, there have been 40 legal actions in 37 states. The EPA is
under court order or consent decree in many states to ensure that TMDLs are established,
either by the state or by the EPA (EPA 2003). This timeline of events with respect to

Michigan is summarized in Figure 1.

 

14 TMDL lawsuit information is available from the EPA at http://www.epa.gov/OWOW/tmdl/lawsuithtml.

15 See hgpzl/wwwepa.gov/owow/tmdl/lawsuit1.htrnl for a summary of court decrees and consent decrees
with implementation timetables ranging from 5 to 20 years; there is one case where no schedule has been
speciﬁed (for Alaska). This list was last updated on 21 October 2003.

24

These earlier violations were procedural, and the courts were able to avoid the
issue of the content of the TMDLs themselves. This avoidance has begun to change,
however, as more recent lawsuits have begun to question the substance of TMDLs. Under
Section 303 and EPA regulations, a TMDL must include the sum of point source waste
load allocations and nonpoint source load allocations plus a margin for error for

uncertainty and a margin for future growth.

Figure 1: Implementation of Total Maximum Daily Loads

 

 

Total Maximum Daily Load specified in
Clean Water Act, Section 303(d)
1972

TMDL not implemented by regulators

 

 

 

Citizen groups ﬁle lawsuits, beginning in
the 1980s to present

 

 

First court order handed down requiring EPA
to establish TMDL plans (Alaska)
1 992

 

 

 

 

H

 

Michigan begins to actively implement
TMDLs in order to avoid lawsuits
1995-96

 

 

 

 

 

 

Source: US. EPA website, various pages.

Many states, however, have not incorporated nonpoint sources in their TMDLs. For
example, New York and Louisiana were sued for only focusing on point sources and
ignoring nonpoint sources. In Louisiana, load allocations from nonpoint sources were

never accounted for, even in cases where they were “inescapably” the dominant source of

25

water quality degradation (Houck 1999). More citizen lawsuits can be expected over the

contents and implementation of TMDLs.l6

The EPA ﬁrst began to take its TMDL responsibilities in earnest in 1991. In April
of that year, it issued a document entitled “Guidance for Water Quality-Based Decisions:
The T MDL Process.” Further pressured by lawsuits, the EPA began to take TMDLs much
more seriously, and issued two guidance documents in late 1997. The ﬁrst document
clariﬁed the scope of listed waters, and for the ﬁrst time, schedules for load allocations
from nonpoint sources, and their contents were speciﬁed in the second document. In
October 1997, the EPA issued a memorandum to its Regional Administrators entitled
“New Policies for Establishing and Implementing T MDLs”, which for the ﬁrst time

speciﬁed its enforcement options to secure state compliance in print (Houck 1999).

With these guidance documents, the EPA speciﬁed three things: 1) a schedule for
the development of all TMDLs by states and a deadline, 2) a focus on waters polluted
primarily or exclusively by nonpoint sources, and 3) the tools that it would use to enforce

implementation of nonpoint source TMDLs.17

2.5 The Total Maximum Daily Load Process in Michigan

Thus, although the TMDL process was speciﬁed under Section 303(d) in the 1972
amendments, it was not enforced since the Clean Water Act has primarily been used to

control point sources of pollution — not nonpoint sources —— over the last 30 years. In

 

16 As an indication that the environmental community will continue to pressure TMDL implementation,
see National Wildlife Federation guide, entitled “Saving our Watersheds: A National Wildlife Federation
Field Guide to Watershed Restoration Using TMDLs”, at
http://www.nwf.org/watersheds/ﬁeldguide/about.htrnl.

‘7 The EPA’s main TMDL website is http://www.epa.gov/owow/tmdl.

26

response to the EPA requirement that the TMDL process to come into effect by October
2001, the Michigan Department of Environmental Quality (MDEQ) developed
Michigan’s TMDL implementation schedule (began in 2000, ends in 2011).18 The

MDEQ was delegated authority from the EPA to implement TMDLs in Michigan.

Section 303(d) requires states to identify and rank water bodies for which
technology—based efﬂuent limitations have not been successful in reaching or maintaining
water quality standards. These water quality standards have been set by the Michigan
Department of Environmental Quality (MDEQ) to protect surface waters based upon
intended uses, that is, recreational, public water supply, agricultural, or industrial uses.
The list of impaired waters is known as the “303 (d) list” or the “dirty waters” list. States
must then develop a TMDL plan or process for each of these impaired water bodies so

that they attain their state-designated water quality standard. 19

The TMDL process requires states to do a number of things: 1) set the maximum
amount of pollution that a water body can receive without violating state-designated
water quality standards (including a margin of safety to account for technical

uncertainties); 2) develop a quantitative assessment of water quality problems, pollution

 

18 A Michigan pilot program is underway in Lake Macatawa in Ottawa County. The Michigan Department
of Environmental Quality (MDEQ) received a grant from the EPA in October 1996 to develop a
phosphorus TMDL for Lake Macatawa. Phosphorus is the limiting nutrient in Lake Macatawa and nearly
always the nutrient which controls the eutrophication level of lakes in Michigan. See
http://www.deq.state.mi.us/documents/deq-swq-gleas-tmdlmacatawapdf.

19The Year 2002 303(d) report for Michigan provides an updated list of Michigan water bodies that are
threatened or do not meet Michigan's Water Quality Standards and require a Total Maximum Daily Load
(TMDL). It is available at mzﬂwwwdeqstatemi.usjdocuments/deq-swq;gleas-303 d Rpt2002b.pdf. The
Michigan Department of Environmental Quality’s web site that deals with TMDLs is at
hgpj/wwwmichigan.gov/deg/O,1607,7-135-3313 3686 3728-12464--,00.htm1. The latest Section 303(d)
list fact sheet for Michigan that summarizes Michigan’s TMDL program and the impaired waters is for the
year 2000 and is available at: http://cE§pub.epa.gov/waters/state_rept.contrcﬂ)_state=MI. Also see the
Elton R. Smith web site for discussion of TMDLs and an EPA-Universities Workshop on TMDLs:
http://www.aec.msu.edu/agecon/smith_endowment/alternatehome.htm.

 

27

sources, and required pollutant reductions; and 3) address all pollution sources — both
point sources such as factories and municipal plants, and nonpoint sources such as runoff
from agricultural lands, forests, and roads. States are given ﬂexibility in allocating the
required load reductions amongst sources. For instance, at the extreme, states could make
point sources responsible for the entire required load reduction in a watershed; or, in
implementing TMDLs, states can choose to mandate technologies for ﬁrms, even though

technologies are not required by the Clean Water Act.

Consequently, TMDLs can be implemented as ﬂexible performance standard
since the legislation does not specify the technology that must be used by farms of a
certain size. In the case of non-CAFO dairies - that is, dairies with less than 700 milking
cows — a TMDL could be speciﬁed as a ﬂexible ambient or efﬂuent environmental
performance standard. If such standards were to be used, innovation offsets may result
within agriculture. Thus, potentially TMDLs provide an opportunity for ﬂexible

environmental performance standards to be applied to agriculture.

Nevertheless, the cost-effective use of ﬂexible environmental performance
standards may be hampered by uncertainties and resulting transaction costs. It is this
component of performance standards that is addressed next in Chapter 3, Uncertainty and

Transaction Costs.

28

CHAPTER 3

Uncertainty and Transaction Costs

The implementation and management of public policies involves transaction costs
(Grifﬁn and Bromley 1982). Whether ﬂexible environmental performance standards are
actually empirically superior to technology-based agro-environmental policies depends
upon the effects of uncertainty and resulting transaction costs. Regulators and farmers do
not have perfect information about current conditions or how the future will unfold; this
lack of perfect information, or uncertainty, creates transaction costs; indeed, the level of
uncertainty affects the magnitude of transaction costs in agro-environmental policies

(McCann and Easter 2000).

Agro-environmental policy design will affect who bears the transaction costs -
the regulator or the farmer. The cost burden is based upon the property rights structure
that assigns responsibilities for environmental outcomes (Batie and Ervin 1999). Flexible
environmental performance standards can be implemented so that either the regulator
bears the bulk of the transaction costs, or the farmer, or the costs could be shared by these
two decision makers. Different institutional structures — or the implementation details of
the regulation — result in differential incidence of transaction costs (Bromley 1986). Since
the incidence of policy costs will determine the political feasibility of ﬂexible

performance standards, the manner in which the policy is designed matters.

3.1 Transaction Costs and Agro-Environmental Policy

Transaction costs were ﬁrst related to environmental policy by Coase in “The
Problem of Social Choice” (Coase 1960); Coase argued that pollution abatement costs

and transaction costs should be considered when evaluating alternative environmental

29

policies. Transaction costs can be deﬁned as the costs of coordinating and motivating
economic agents (Milgrom and Roberts 1992). The coordination problem in agro-
environmental policy is to determine what pollutants should be reduced, how they should
be minimized, and who should undertake appropriate actions. The motivation problem is
to ensure that farmers willingly do their parts, both in reporting information accurately to
allow the best policy to be designed, and to act as speciﬁed by the policy (Milgrom and

Roberts 1992).

McCann and Easter (1999) state that transaction costs (including administrative
costs) may be particularly high for nonpoint source pollution due to the high cost of
monitoring efﬂuent from numerous polluters. These higher costs are a contributing factor
as to why point sources historically have been emphasized in water quality legislation as
opposed to nonpoint source pollution. In the context of agro-environmental policy for
nonpoint source pollution, an examination of transaction costs is important because they
may affect which policy alternative attains an environmental goal at least cost, and their
examination may lead to the design of policies and institutional arrangements which
lower some transaction costs (McCann and Easter 1999). Aside from uncertainty (Section
1.2.1), there are two other sources of transaction costs that are relevant for this study of
nonpoint source pollution agro-environmental policies, namely bounded rationality and

opportunistic behavior.

3.1.1 Sources of Transaction Costs: Bounded Rationality

Bounded rationality describes the limited capacity people have to process
information. As Simon illustrated, the mind is a scarce resource (Simon 1978). Thus,

economic agents — both regulators and farmers -— are rational, but only in a limited

30

 

 

fashion; they only use a subset of the information available to them when making
decisions. Individuals therefore often make decisions based upon incomplete knowledge
of all possible alternatives and their likely outcomes (Ostrom, Schroeder, and Wayne
1993). Limited foresight and the costs of calculating solutions help to explain the

existence of bounded rationality.

Bounded rationality helps to explain why farmers do not pursue all proﬁtable
opportunities for new production processes.20 The set of information considered by
farmers is inﬂuenced by such factors as market prices, institutional factors such as
business customs and culture, social culture, and property rights. Changes in any of these
institutions or market prices — perhaps due to citizen movements, political responses, or
changes in property rights — can cause farmers to widen their information set and change
their optimizing behavior; for example farmers may respond by searching for innovative

production processes that reduce water pollution.

On the other hand, due to bounded rationality, they may not. Even though there
may be proﬁtable paths that farmers can pursue now within a new institutional
environment, they may explicitly choose not to do so due to other concerns, for instance,
due to a certainty payback threshold. An example of a payback threshold may be that the
farmer wants to earn, say, six per cent above his costs per year, and will only consider
those enterprises that can guarantee that level of revenue. Simple measures like certainty
payback thresholds help farmers focus their attention by allowing them to deal with a

subset of their entire knowledge. Bounded rationality thus can affect innovation choices.

 

20 This section is adapted from Srivastava, Batie, and Norris (1999).

31

Regulators also are subject to bounded rationality; they may not pursue all
signiﬁcant optimal opportunities to improve environmental outcomes. The set of
information considered by regulators is inﬂuenced by such factors as market prices,
institutional factors such as regulatory customs and culture, social culture, property
rights, and political pressures. As with farmers, changes in any of these institutions or
market prices may force regulators to change their information set and their optimizing

behavior.

3.1.2 Sources of Transaction Costs: Opportunistic Behavior

With complicated issues such as nonpoint source pollution, another source of
transaction costs is opportunistic behavior. Opportunism is the deceitful behavior
intended to improve one’s own welfare at the expense of others (Ostrom, Schroeder, and
Wynne 1993). Due to informational asymmetry — in this case, the regulator does not
possess perfect information with respect to farmers’ production practices and behavior —
farmers have an informational advantage. They may take advantage of this informational

asymmetry and behave strategically, or opportunistically.

When those with critical information have interests different from those of the
policy maker, they may fail to report completely and accurately the information needed to
make decisions (Milgrom and Roberts 1992). For example, a farmer may, strategically,
fail to disclose aspects of his management practices to regulators, or he may disguise or
even distort information regulators need to effectively control nonpoint source pollution.
This opportunistic behavior leads to the problem of moral hazard. Moral hazard arises
when one agent, here the regulator, is imperfectly informed of the actions of another

agent, here the farmer. If the regulator can not ascertain which farmer is generating

32

pollution and which farmer is not — and farmers know that this is the situation — some

farmers may choose to behave opportunistically and not comply with the TMDL.

The regulator does possess some information about the nonpoint source pollution
problem, such as an estimate of the fate and transport of pollutants or the number of
farmers in a watershed. The regulator, however, does not possess detailed, perfect
information regarding each farmer’s production process. Conceptually, if the regulator’s

set of information is denoted Q, and the farmer’s is denoted Q f , then due to information
asymmetry, these information sets will not be equal, that is, (2, ¢ (2] . Moral hazard. thus
implies that if the farmer’s production process is represented by a production function

f (it) , where )‘c is the vector of inputs that the farmer chooses, the regulator does not
know what f (it) is, although the regulator does form expectations of f (56) .
Consequently, if the regulator’s expectation of the farmer’s production process is f, (5?, ) ,
then fr (f,) at f (it) — either because the regulator does not know what the functional

form is, and/or the regulator does not know the farmer’s input vector. The unequal

information sets allow the farmer to behave opportunistically.

The possibility of opportunistic behavior by some farmers increases the
transactions costs of agro-environmental policies not just for regulators, but other farmers
as well. Regulators must design agro-environmental policies to provide incentives that
reward farmers while simultaneously constraining their behavior in such a manner that
they internalize the social costs that they generate; that is, the agro-environmental
policies must align social and private objectives. This alignment is no easy task when

farmers are able to behave opportunistically. It is unlikely that all farmers will behave

33

opportunistically; some will comply with the requirements of the agro-environmental
policies and will reduce water pollution from their operation. Some farmers, however,
will choose to behave opportunistically. It is due to this latter group that monitoring and
enforcement measures must be implemented; thus, farmers who do comply have to
endure monitoring and enforcement hurdles since the regulator can not easily distinguish

one group from the other.

3.2 Types of Transaction Costs

Williamson (1985) identiﬁed two types of transaction costs, ex ante transaction
costs, and ex post transaction costs. Ex ante transaction costs are largely co-ordination
costs that exist to a greater or lesser extent in long-term economic relationships, like
those that arise from agro-environmental policies (Ostrom, Schroeder, and Wynne 1993).
Since unforeseen situations arise and all future contingencies can not be envisioned,
transaction costs also occur in long-term economic relationships that involve multiple
agents, for instance, after an agro-environmental policy is implemented (Ostrom,
Schroeder, and Wynne 1993). These costs are ex post transaction costs. Thompson
categorizes transaction costs for environmental policies into six areas: 1) research and
information, 2) enactment, 3) design and implementation, 4) support and administration,

5) monitoring, and 6) prosecution (Thompson 1999).

3.2.1 Transaction Costs Incurred by Regulator

The regulator can incur both ex ante and ex post costs when implementing a
ﬂexible ambient environmental performance standard due to a lack of information.

Research and information, enactment, design and implementation, and support and

34

administration are types of ex ante transaction costs. The ex ante costs that the regulator
could incur are outlined in Table 1; selected costs associated with a TMDL are used as an

illustration.

In order to implement the TMDL process, the regulator must gather biophysical
data to ascertain the appropriate level of phosphorus for a speciﬁed body of water, given
its designated water uses; then the regulator must process that information in order to
determine the allowed ambient level of phosphorus. Finally, the regulator must decide
how to allocate the allowed polluted efﬂuent amongst the dischargers into the body of

water.

Table 1: Selected Ex ante Transaction Costs Incurred by Regulator

 

 

Data Gathering Costs Data Processing Costs Decision Costs
- Data to determine appropriate - Determination of appropriate - Determination of
ambient phosphorus level ambient phosphorus level hOW TMDL Will be
- Data to determine actual - Determination of actual allocated amongst
phosphorus level phosphorus level dischargers
- Data to determine who is - Determination of who is - Evaluating
discharging, and their efﬂuent discharging, and their efﬂuent performance
levels levels, or monitoring activities

Determination of
penalty level

 

Some of these ex ante transaction costs can be high with respect to nonpoint
source pollution. Of special interest are the information search costs 4 the data gathering
and processing costs that the regulator must bear in order to determine individual
discharge sources’ efﬂuent levels, since the diffuse nature of nonpoint source pollution
makes it difﬁcult to determine the sources, let alone how much efﬂuent each discharger is
releasing. In response to these ex ante transaction costs caused by uncertainty, the

regulator may implement ﬂexible ambient performance standards in a manner that

35

reduces these costs, or it may pursue a command-and-control strategy such as requiring
the adoption of on-farm technologies that control or prevent pollution. The cost-savings
from these differing compliance measures are explored in detail using a simple

conceptual model in Chapter 6.

Similarly, Table 2 illustrates selected ex post transaction costs that the regulator
might incur. Under Thompson’s categories, ex post costs include implementation, support
and administration, monitoring, and prosecution. With the TMDL example, the regulator
must monitor the'water body to ensure that the speciﬁed phosphorus level is maintained,
again incurring information search costs; additionally, the regulator must enforce any
required reductions and assess relevant penalties. If subsidies are used as an incentive
measure, then they must also be assigned. Since the regulator does not know the
production function of each farmer, and can only form expectations of the production
process, it can not easily determine the efﬂuent from each farm. Due to this uncertainty
generated by nonpoint source pollution and the possibility for farmers to behave
opportunistically, regulators will have difﬁculty in determining who is in compliance,

resulting in ex post transaction costs.

Table 2: Selected Ex post Transaction Costs Incurred by Regulator

 

 

Monitoring Costs Enforcement Costs
- Monitoring ambient phosphorus level - Enforcing required reductions
- Monitoring efﬂuent from discharge sources - Assessing penalties/subsidies

- Determining who is or is not in compliance

 

In fact, this uncertainty results in large ex post costs for the regulator, both in
terms of monitoring and enforcement costs. Nonpoint source pollution implies numerous,

diffuse sources of nutrient efﬂuent or loadings, making it almost impossible to perfectly

36

monitor behavior and enforce the ambient pollution level (Hanley et al. 1997). Moral
hazard and opportunistic behavior result because the information sets of the regulator and

the farmer are not the same, Q, st Q] , and the regulator only can form expectations

regarding a farmer’s production function, which will not be accurate ( f, (fr) :2 f (x‘) ).

Where moral hazard exists, where costs of compliance are great, and where
farmers believe they can free-ride, some farmers may decide to avoid any abatement
effort. In fact, in order to “perfectly enforce” the TMDL in such a situation, in-ﬁeld
pollution “police” would have to be stationed permanently at all ﬁelds that received
manure applications (Innes 1999). But, such “perfect enforcement” is obviously neither
politically nor ﬁnancially feasible. As a consequence, the regulator may wish to reduce
these signiﬁcant ex post transaction costs by increasing compliance certainty. For
example, the regulator could implement the TMDL as an efﬂuent standard or a design

standard instead of an ambient standard.

These high ex ante and ex post transaction costs imply that that regulator may be
unable to effectively implement the TMDL process, or any ﬂexible ambient performance
standard, potentially resulting in failed environmental outcomes such as diminished water
quality. If water quality does not improve due to poor TMDL implementation, the
regulator becomes vulnerable to lawsuits from citizen groups for failing to implement the
Clean Water Act. Grifﬁn and Bromley (1982) recognized that litigation costs must be
considered along with monitoring and assessment in policy transaction costs. The
possibility of lawsuits against the agency may cause the regulator to seek ways to

minimize compliance uncertainty when implementing the TMDL process. Consequently,

37

mandating and enforcing strict environmental compliance with a ﬂexible ambient

performance standard may not be possible with nonpoint source pollution.21

3.2.2 Transaction Costs Incurred by Farmers

Similarly, farmers also face uncertainty with respect to the pollution that they
generate and compliance with the total maximum daily load process that is implemented
as a ﬂexible ambient standard.22 Unlike the regulator, however, they do not incur any ex
ante transaction costs. Farmers react to the ﬂexible ambient standard —— thus they only
incur transaction costs after TMDLs are in place. Until the TMDL is implemented, water
degradation has no pecuniary value, so farmers are unlikely to consider most of the

degradation costs associated with their production process.23

Once the TMDL is in place, farmers incur ex post transaction costs, due to
compliance uncertainty. It is difﬁcult, and hence costly, for farmers to link the effects of
their management practices to runoff from their lands. For instance, even if farmers
spread manure onto their lands at a rate below the upper limit speciﬁed in Michigan
animal waste guidelines, a severe storm can result in signiﬁcant soil erosion into the

water body.24 Phosphorus binds to soils, so if enough soil is washed into the water body,

 

21 The Porter Hypothesis maintains that strict enforcement of a binding standard is necessary to encourage
technological innovations.

22 Farmers may also face uncertainty by a ﬂexible efﬂuent standard, depending upon how much
information the regulator shares with the farmer.

23 This argument is not to say that the farmer will not care about any of the social costs of the water
degradation; he may well consider the effects of any water degradation that he causes even though he does
not directly pay for the degradation. He may consider these effects due to his own ethical beliefs, or due to
legal liability concerns.

24 Under the Michigan Right to Farm Act (1981 ), the Michigan Commission of Agriculture develops and
adopts Generally Accepted Agricultural and Management Practices (GAAMPs) for farms in Michigan.
These are voluntary practices meant to protect farmers from nuisance suits. The Manure Management and
Utilization GAAMPs can be viewed at:
http://wwwmichigan.gov/documents/MDA_manureRedlineStrikeoutZ003_676 l 5_7.pdf.

38

the phosphorus limit enforced by the TMDL process may be exceeded. The level of these
costs depends upon how the regulation is implemented. For instance, the regulating
agency could allow the standard to be exceeded for some portion of the year, to allow for
extreme weather events, or the regulating agency or the government could provide

ﬁnancial and or technical assistance to the farmer to reduce these transaction costs.

Since the total maximum daily load process is speciﬁed as an ambient standard,
the penalty system may be based upon this ambient standard.25 But with a ﬂexible
ambient environmental performance standard, even if livestock farmers make expensive
investments in manure management systems, regulators will ﬁnd it difﬁcult to distinguish

the effects of each farmers’ actions ( f, (2,) i f (x) for each farmer), so an individual

farmer may still be found to be out of compliance due to the aggregate efﬂuents from a

group of farmers, regardless of his own efforts.

This informational deﬁciency could reduce the chances that a farmer will comply
with a total maximum daily load unless the expected penalty for noncompliance is very
high. Some farmers may choose to not comply with the ﬂexible ambient performance
standard because they may be tempted to free ride — to let others undertake pollution
abatement measures while they share in the beneﬁts of reduced nutrient loadings in the
body of water. These farmers may choose to free ride due to the compliance uncertainty
that exists for regulators. As a consequence, in the absence of a high expected penalty for
non-compliance, signiﬁcant compliance uncertainty may dissuade farmers from making

investments in innovative technology. Thus, it would appear that strict compliance by

 

25 States have the ﬂexibility to convert this ambient-based penalty system to another type, such as an
efﬂuent system.

39

farmers may not be possible with a ﬂexible ambient environmental performance standard

targeted at nonpoint source pollution.

If, however, the regulator could perfectly monitor and enforce the ambient
standard embedded as a TMDL, the regulator would know perfectly each farmers’
actions and the environmental effects of each farmers’ actions. Conceptually, this
situation would be as if its expectation of the farmer’s production function coincides with

the farmer’s actual production function, so f, (56,) = f (i). In this case, moral hazard

would not be an issue, and depending upon penalties and compliance costs, the incentive
for farmers to behave opportunistically would decrease, as would the motivation to free
ride. With perfect monitoring and enforcement, compliance uncertainty for regulators is
eliminated, and farmers have greater motivation to comply since their efforts would be

correctly recognized and they would not be wrongfully penalized.

Even without perfect monitoring and enforcement, some farmers may be
interested in complying with nonpoint source policies due to recent legal battles in the
livestock industry, if compliance uncertainty is reduced. Under the Clean Water Act, a
third-party can sue a farmer or any other ﬁrm for violating the Clean Water Act. This
situation is particularly relevant for Michigan, where in the summer of 2000, the Sierra
Club gave “notice to sue” to ﬁve dairies in the state, effectively increasing those farmers’
expected penalty for noncompliance. If compliance uncertainty is eliminated, so that it is
easy for regulators and third parties to determine who is in compliance with a ﬂexible
ambient environmental performance standard, then some farmers will be interested in
complying in order to reduce the likelihood of being sued. Farmers thus may seek ways

to increase compliance certainty.

4O

3.3 Flexibility and Compliance Certainty Trade-off

Thus, “ﬂexibility” in ﬂexible environmental performance standards contributes to
compliance uncertainty and this compliance uncertainty, in turn, results in transaction
costs.26 Contrast this case with a design standard, where both the regulator and farmer
can easily determine compliance; the regulator simply checks whether the farmer has
adopted the speciﬁed technologies, and the farmer knows he is in compliance as long as
he implemented the required technologies. Design standards, however, may not correlate
well with improved environmental conditions. There appears to be trade-offs between
ﬂexibility and compliance certainty embodied in those environmental performance
standards that address nonpoint source pollution. From a policy point of view, Michigan
regulators need to decide how much ﬂexibility to embed into the TMDL process to
control phosphorus pollution from dairy farms in the state, since the ﬂexibility and
certainty characteristics will affect the success of this approach, and ultimately the

policy’s ability to encourage induced innovation and innovation offsets.

One way to analyze this trade-off when regulating nonpoint source pollution with
a TMDL strategy is via a production possibilities frontier (PPF), illustrated in Figure 2.
The frontier curve shows the various combinations of compliance certainty and ﬂexibility
that can be provided by, or “produced” by an environmental regulation. The inputs are
employed, or allocated, efﬁciently. Compliance certainty is measured on the X-axis; it
increases as one moves from left to right. Compliance refers to farmer compliance with
the legislation as implemented through mandatory regulations. Flexibility — on the Y-axis

— is deﬁned as the ability of the farmer to be free to determine how to become compliant

 

26 The stochastic nature of nonpoint sources of pollution and moral hazard also contribute to this
uncertainty.

41

 

 

with the regulations. The production possibilities frontier is downward-sloping because
of the trade-off between ﬂexibility and compliance certainty; there is an opportunity cost
in producing one good over the other. Additionally, it is assumed that there are
specialized inputs in the production of each good and differing factor intensities for
common inputs. Thus, for an environmental legislation to “generate” — or allow —— more

compliance certainty, the level of ﬂexibility produced — or allowed — by it declines.

Figure 2: Flexibility and Compliance Uncertainty Frontier

 

   
 
   
    

Flexibility “ Flexible Ambient Environmental Performance
Standard
A
Fl

Flexible Efﬂuent
Environmental Performance
Standard

Design Standard-based
Environmental Regulation

 

 

 

 

0 CI c,

 

Compliance Certainty

 

 

Note that this trade-off may not exist, but only in the extreme case of perfect
monitoring and enforcement by the regulator. In this situation, the regulator would know
the exact effect of the inputs and technology used by the farmer on water quality, or
f, (7g) = f (56). While the farmer has the ﬂexibility to meet the standard in the most cost-
effective manner for his farm operation, the regulator incurs an extremely high cost of

information gathering. This avenue is prohibitively expensive, however. In the case of

42

 

imperfect monitoring and enforcement, there will be a trade-off between ﬂexibility and

compliance certainty.

Consider Figure 2. As one moves north along the Y-axis, the amount of ﬂexibility
produced by the environmental regulation increases, while compliance certainty
decreases. As such, point A corresponds to a ﬂexible ambient environmental performance
standard since the amount of ﬂexibility allowed is quite high — F1 — and it results in a
relatively low level of compliance certainty, C]. At point A, regulators can take
advantage of this regulatory ﬂexibility by accommodating diverse natural resource
conditions and heterogeneous production systems; additionally, ﬂexible environmental
regulations remove the onus on regulators to determine the “best” technology to be used
by farmers. For farmers, the ﬂexibility provides dynamic incentives for technological

innovation and cost-effectiveness through possible innovation offsets.

Point B on the frontier corresponds with an environmental regulation that uses a
design standard with a greatly diminished level of ﬂexibility (F2) but a much higher level
of compliance certainty, C2. At point B, some compliance uncertainty is removed so
some transaction costs decline, but at the cost of reduced ﬂexibility for farmers. Also, the
regulator must now specify the approved technologies. Despite the diminished level of
ﬂexibility at this point, and less certainty with respect to resulting ambient water quality
improvements, both regulators and farmers may prefer point B if enough of their

transaction costs are sufﬁciently reduced.

A third point is depicted on the frontier, labeled Flexible Eﬁluent Environmental
Performance Standard. With this type of standard, the regulator speciﬁes some level of

efﬂuent that can leave the farmer’s operation, but the farmer is free to choose whatever

43

 

 

technology he wants to meet his efﬂuent performance standard. Consequently, this case
lies between Points A and B in the amount of certainty and ﬂexibility produced by this
type of compliance measure or environmental regulation; there is greater certainty — and
no ﬂexibility — in terms of what is expected of the farmer, but he still has ﬂexibility — and

uncertainty — in how he achieves his efﬂuent standard.

If the regulator is willing to trade—off ﬂexibility in order to reduce some of its ex
ante and ex post transaction costs through an increase in compliance certainty. If point A
represents a ﬂexible ambient standard — incorporating a lot of ﬂexibility but no certainty
— and point B represents a design standard — incorporating a lot of certainty but no
ﬂexibility - the regulator may choose to choose to move from point A to B and convert
the ﬂexible ambient standard into a design standard. By dictating to farmers what
technology to use, to an extent, the regulator has increased information to farmers;
institutional arrangements that help to generate or distribute information serve a crucial

role in reducing some transaction costs (Ostrom, Schroeder, and Wynne 1993).

In fact, some farmers may prefer point B to point A due the ex post transaction
costs that they incur and because they gain compliance certainty. At point A, they have to
bear information costs in seeking out a technology that will allow them to reduce their
pollution loadings or efﬂuent. Additionally, due to compliance uncertainty, a farmer may
be mistakenly penalized and/or sued for water degradation despite the technology that
they have sought and implemented, since it is difﬁcult for the regulator to link
management practices to environmental performance with nonpoint source pollution and
an ambient standard. Even if farmers would like to undertake abatement efforts, liability

issues may prevent them from innovating.

44

 

 

 

In such a situation, if farmers face liability issues that are not due to their own
efforts, they may prefer to shift liability costs to the regulator by demanding design
standards speciﬁed by the regulator. By then following the design standards, farmers may
be able to reduce their compliance costs, without having to be innovative. Consequently,
some farmers may feel that their information search transaction costs are not worth
incurring, and thus may prefer the greater level of compliance certainty provided by a
point like B. The transaction costs associate with a ﬂexible ambient standard may affect
the beneﬁts that farmers could receive through induced innovation, and thus affect the

likelihood that an innovation could occur.

Which point will regulators and farmers prefer? Will compliance certainty be
more important or ﬂexibility for each as they contend with total maximum daily loads?
Whether regulators will prefer to shift the informational transaction costs of determining
the best technology to use onto farmers through ﬂexible performance standards, or
whether they would rather prescribe design standards and increase some compliance
certainty is unknown. Similarly, whether farmers will embrace the ﬂexibility aspect of
ﬂexible performance standards and seek out least—cost pollution abatement technologies,
thereby beneﬁting from possible innovation offsets is unknown in the presence of
compliance uncertainty since they also have to contend with possible lawsuits. The
objectives of each agent will determine whether induced innovation and innovation

offsets will occur, as will the relevance of the Porter Hypothesis in the face of compliance

uncertainty.

If the Porter Hypothesis is relevant for agriculture, then the political feasibility of

TMDLs may be enhanced since environmental quality improvements are achieved in

45

 

conjunction with some beneﬁts to farmers through possible innovation offsets. F arrners
become an integral part of the overall strategy to resolve agro-environmental problems
once they have the incentive to maintain a healthy, sustainable, and agriculturally
productive Great Lakes region. A closer examination of the Porter Hypothesis and
innovation offsets will provide insights into whether innovation offsets are possible when
total maximum daily loads are used to control nonpoint source pollution in the face of
compliance uncertainty. The Porter Hypothesis and its application to nonpoint pollution

is the subject of Chapter 4.

46

 

 

CHAPTER 4

Porter Hypothesis and Innovation Offsets

The Porter Hypothesis has generated considerable debate. Porter and van der
Linde contend that innovation offsets are likely to be common and large, while others
have disagreed (Gardiner and Portney 1994; Palmer, Oates, and Portney 1995; Thurow
and Holt 1997; Jaffe and Palmer 1997). For example, Palmer et al. (1995) note their
strong dispute with the Porter Hypothesis; they argue that environmental regulation does
indeed involve trade-offs, and that the cost of regulation will be neither negligible nor
non-existent. Which of these arguments will be true in the case of TMDL standards being
applied to control nonpoint source pollution from dairy farmers? Will innovation offsets
occur when ﬂexible environmental performance standards are used to manage nonpoint

source pollution? These issues are examined and clariﬁed below.

4.1 Behavior Implications

How will farmers react to a new ﬂexible ambient environmental performance
standard? Can farmers offset decreased proﬁts which stem from compliance with a new
TMDL water quality requirement? Hicks’ induced innovation hypothesis provides

theoretical insight:

A change in relative prices of the factors of production is itself a sign to
invention, and to invention of a particular kind -- directed to economizing

the use of a factor which has become relatively expensive (Hicks 1932).

The change in property rights due to the new water quality requirements has
caused the relative cost of water to increase for farmers. It follows from Hicks’ induced

innovation hypothesis that a farmer will choose a technology that changes his production

47

process and reduces the amount of water that his farm uses for pollution disposal. A
proﬁt-maximizing farmer has an incentive to invest in ways to meet his new constraint at

lower costs.

Is Hicks’ induced innovation hypothesis — which focuses on inputs — relevant for
an environmental performance standard, an output-based measurement? Yes, because
water quality is affected by the production process; that is, through his production
process, the farmer creates joint goods. For a dairy farmer, the joint goods are milk and
water pollution. Before an ambient environmental performance standard is implemented,
water quality has no value or cost to the farmer, so using a lake or a river as a sink for
nonpoint source pollution is costless or free for him. The imposition of the ﬂexible
ambient environmental performance standard, however, causes water pollution to become
a cost for the farmer; in other words, water is no longer a free input into the farmer’s
production function. This cost is a portion of the social cost of water pollution that the
farmer creates and is now being forced to internalize due to the ambient environmental
performance standard.27 In essence, water quality is like an input that now has a cost for
the farmer. Hicks’ induced innovation hypothesis states the farmer will substitute away

from using water for waste disposal.

The implications of Hicks’ induced innovation hypothesis, however, is not so
straightforward in the case of nonpoint source pollution. In this situation, the farmer’s
action in response to the environmental regulation may not result in innovation offsets.

With nonpoint source pollution, the occurrence of innovation offsets depends upon the

 

27 If the ambient performance standard is set at the socially optimal level, then the farmer intemalizes the
full social cost of the water degradation that he causes. How much of this internalized cost is ultimately
borne by the dairy product consumer and how much is borne by the dairy farmer will be determined by the
elasticities of supply and demand.

48

effect of compliance uncertainty. Due to moral hazard and compliance uncertainty, it is
difﬁcult to link investments and abatement efforts with positive environmental outcomes,
thus a farmer may not be given credit for economizing on his use of water as a disposal
sink; without this beneﬁt, and in the absence of a sufﬁciently high, certain penalty, the
farmer may not seek out innovative ways to reduce water pollution, thereby precluding

innovation offsets.

Instead, farmers may ask for a list of speciﬁc approved technologies from the
regulator to reduce their compliance uncertainty-related transaction costs. If the regulator
also wishes to increase certainty and reduce some of its compliance uncertainty
transaction costs, then the ﬂexible environmental performance standard may be converted
into a design standard. Since speciﬁed technologies preclude induced innovation and
innovation offsets, the question is: will compliance uncertainty offset the incentive for
farmers to innovate and thus prevent innovation offsets from occurring? Will innovation
offsets be a straightforward result of strict environmental regulations as Porter and Van
der Linde assert? Or will they be negligible? Do trade-offs exist? A graphical model is

presented next to help answer these questions.

4.2 A Simple Model of Innovation Offsets

Much of the debate over the Porter Hypothesis is due to a lack of clarity regarding
the baseline to be used in determining whether innovation offsets can occur as a result of
environmental regulations. The following simple model clariﬁes this question and

provides a straightforward deﬁnition of innovation offsets.28

 

28Adapted from Batie 1997.

49

With the implementation of the environmental regulation, and the change in
property rights, pollution is now a cost for farmers. Economic theory predicts that a
farmer will reduce his use of water in order to minimize costs due to the new ﬂexible
ambient environmental performance standard. Figure 3 illustrates the initial effect of this
change in property rights. The production possibility frontier (PPF) depicts the feasible

set of water quality and proﬁts that are produced by a livestock farm, say Farm A, with a

 

 

 

 

 

 

 

given technology.
Figure 3: Proﬁts and Environmental Quality Frontier
Proﬁt
A0
P .\?Al\
1
P, A2
XPPFI g
0 W. W? Environmental Quality

 

 

 

Adapted from Srivastava, Batie, and Norris (1999).

The production possibility frontier is used because it is an important tool for
analyzing the simultaneous production of two outputs. The curve demonstrates that there
are many possible allocatively efﬁcient combinations of proﬁt and water quality and that
producing more of one necessitates a reduction in the production of the other; in short,

there is an opportunity cost to each point on the frontier. In Figure 3, environmental

50

quality is water quality, and is measured on the X-axis; it increases as one moves from

left to right. Firm proﬁt is on the Y-axis, and increases as one moves north.

Inputs are characterized to be allocatively efﬁcient when the only way to increase
the output of one commodity is to decrease the output of the other. The frontier is the
locus of allocatively efﬁcient points and implicitly deﬁnes the relationship between
proﬁts and environmental quality. The farmer is assumed to be allocatively efﬁcient and
is thus on the frontier. As in Figure 2, the production possibilities frontier is downward-

sloping because it is assumed that there are differing factor intensities for common inputs.

This production technology describes the method in which inputs are combined to
produce outputs, in this case proﬁts and water quality. This conceptual abstraction of
technology encompasses different processes, capital, and management practices —
essentially different methods of production. Farm A uses the same technology as long as
it is on the same production possibility frontier, and is able to produce different levels of

water quality and proﬁts using this technology.

Before the environmental regulation, Farm A is on some point on the production
possibility frontier labeled PPFi. Economic theory does not predict where on the frontier
Farm A will be; it depends upon the farmer’s utility function. Farm A will be on the point
on the frontier that is tangent with the farmer’s highest utility function. If the only
parameter in the utility function is proﬁt, that is, the farmer is a proﬁt-maximizer, then
Farm A would be at point A0. At this point, no resources are devoted to producing any
positive water quality; consequently, proﬁts are as high as possible with existing
resources. Since environmental quality has no ﬁnancial value for Farm A and using water

as a sink is free for the farm, the proﬁt-maximizing farm would be at A0.

51

But, if the farmer’s objective function is to maximize utility, and he cares about
other issues, such as taking a defensive position due to the threat of lawsuits, or if he is
concerned by the effects of water degradation, Farm A could be at a point like A]. For
illustrative purposes, let Farm A be at A1 where it earns P1 proﬁts and generates a water

quality level equal to W].

Now let a ﬂexible environmental performance standard be implemented (Figure
3) through a TMDL process. Assume the TMDL speciﬁes the maximum amount of
ambient phosphorus that can be allowed from Farm A. Let the minimum level of ambient
water quality that A must meet be Wp. Assume that the regulator is able to strictly and
perfectly enforce the ﬂexible ambient performance standard at the farm level. This

assumption removes the compliance uncertainty for both the regulator and Farm A.

At point A1, it is clear that Farm A is producing too much pollution and must
move to at least A2 along the frontier PPFl. Here Farm A earns P2 proﬁts, which is less
than P1, but it produces less pollution. The difference in P1 and P2 is the cost Farm A
incurs to be in compliance with the new ﬂexible ambient performance standard; that is,

P1 - P2 is Farm A’s compliance cost of meeting the ﬂexible TMDL standard.

If Farm A is able to increase its proﬁts relative to P2 through technological
innovations, then it will be able to beneﬁt from innovation offsets allowed by the

ﬂexibility embodied in the agro-environmental policy.

Farm A, however, is on the same production possibility frontier as before, so this
reduction in proﬁts shown in Figure 3 is not due to any technological change. Rather, it is
the change in property rights that has forced Farm A to use its current technology to

produce a different set of output levels than when it was owner of the property right to

52

pollute. There is a trade-off between proﬁts and improved environmental quality for

Farm A since the TMDL compliance costs have reduced Farm A’s proﬁts.

4.3 Where Will a Farm Be?

If Farm A wants to increase its proﬁts, Hicks’ induced innovation hypothesis says
it will seek out technologies that allow it to economize its use of water. Assume it has
identiﬁed a suitable new technology, so Farm A shifts out to a new frontier. This shift in
the production possibility frontier is an induced innovation. The new perfectly enforced
ﬂexible ambient environmental standard has provided Farm A with the incentive and the
ability to be innovative. Figure 4 illustrates the effect of this induced innovative change

to a new production possibility frontier curve labeled PPF2.29

But where will Farm A be on the new production possibility frontier? Economic
theory does not foretell exactly where Farm A will be on PPF2. It is known, however, that
Farm A must be to the right of Wp because of the ﬂexible ambient performance standard.
There are three possible cases for where Farm A may be on the outer production

possibility frontier.

Case 1 — Below P2: If it is on any part of PPF2 below the horizontal line P2, such
as A3, water quality does improve, but Farm A earns less proﬁts than at A2. At A3, its
compliance costs are greater than they were when Farm A used its original technology
and complied with the ﬂexible ambient environmental performance standard. Although
water quality does improve, Farm A does not lower its compliance costs at points along

this portion of PPF 2; thus, points like A3 are not innovation offsets. It is unlikely, though,

 

29 The technology that is adopted by Farm A could be biased towards proﬁts or environmental quality, in
which case the outer boundary would be closer to the respective axis. For illustrative purposes though, the
production possibility frontier is assumed to take the shape shown.

53

that Farm A will undertake the expense of seeking and employing a new technology only
to earn less proﬁts than it did with the old technology; a proﬁt-maximizing farm will not
end up at a point like A3 which is below P2. A utility-maximizing farmer may be at A3 for
other reasons, but, in that case, he chooses to be there because it is deriving some beneﬁt

that offsets its higher compliance cost.

I

Figure 4: Induced Innovation and Innovation Offsets

 

A

Proﬁt

 

Enhanced Innovation Offsets

 

 

 

 

 

 

0 W: WP Environmental Quality

 

 

 

Adapted from Srivastava, Batie, and Norris (1999).

Case 2 — Between P1 and P2: Assume Farm A moves to A4 on the outer
production possibility frontier. At A4, environmental quality improves relative to A2, as
do proﬁts. After the induced innovation has occurred, at a point like A4, Farm A has
reduced its compliance costs and improved proﬁts relative to A2. Such positions are
deﬁned to be innovation offsets since water quality has improved at the same time that
the farm has managed to reduce its compliance costs due to induced innovation. Farm

A’s proﬁts, however, do not improve relative to A1, where Farm A earns P1 level of

54

proﬁts before the TMDL is enforced. Thus the constraint of the TMDL has reduced Farm
A’s proﬁts, so it is worse off than it was at A1, though others in society are better off

from the improved water quality.

Case 3 — Above P1: Given the way that the production shifts in Figure 4,30 a
proﬁt-maximizing farm will not choose points between P1 and P2, because it can earn a
higher proﬁt and still meet the water requirement at A5. In fact, at points like A4, it is
providing a higher level of water quality than required, Wp while earning a lower level of
proﬁt than it can at A5. Thus, a proﬁt maximizing farm will choose to be at A5. In fact, at
A5, the proﬁt-maximizing Farm A is earning a higher proﬁt than it could at A0, before the
TMDL standard was implemented. Farm A earns this higher level of proﬁt at all points
on PPF2 that are to the right of the vertical line Wp and above the initial proﬁt level. All
points on this section of the production possibility frontier result in higher proﬁts and
improved water quality than at the initial unconstrained point. Points such as A5 are
deﬁned as enhanced innovation oﬂsets since the farm earns higher proﬁts than its initial
situation before implementation of the ﬂexible ambient performance standard while also

improving water quality.

A utility-maximizing farm may choose a point on the enhanced innovation offsets
portion of PPF2, depending upon the relevant utility function. As long as it is on this
portion of the frontier — to the right of Wp and above P. — it will earn a higher proﬁt and
also produce a higher level of water quality than at A1. Thus, a utility-maximizing farm
may also beneﬁt from a perfectly enforced ﬂexible ambient environmental performance

standard.

 

30 It is possible to draw the shift in the production function so that this case is not available to the farmer,
and he must content himself with only innovation offsets which offset the cost of complying with Wp.

55

The ﬂexible ambient performance standard has forced Farm A to Wp, but the
ﬂexibility in the environmental performance standard allows Farm A to reach the outer
production possibility frontier, PPF 2. If the ambient performance standard was not
enforced, a proﬁt-maximizing Farm A could actually be at a point like A6, where it
generates the level of water quality that is most proﬁtable, zero, and earns even a higher
proﬁt than at A5. The difference in proﬁt between A6 and A5 is an opportunity cost that
the ambient performance standard forces Farm A to bear since it is constrained in
meeting the water quality level WP.“ Nevertheless, the perfectly enforced ﬂexible
ambient performance standard provided the incentive to F arm A to innovate, resulting in

the induced innovation. Without the standard, Farm A may not have reached A6.

4.4 Are Innovation Offsets Possible?

Does this mean that Porter and van der Linde were correct and their critics
wrong? Not so. Critics (Gardiner and Portney 1994; Palmer et al. 1995) are correct in
stating that “we cannot have it all.” This conclusion is seen from Figure 3 and Figure 4.
There is an opportunity cost to any position along both frontiers, including PPF2; there is
a trade-off between proﬁts and environmental quality due to scarcity of resources. In fact,
this trade-off will always exist regardless of the technology that is adopted. If the ﬂexible
ambient performance standard did not have to be met, the farm could improve proﬁts at

the cost of lower water quality.

Regardless of this trade-off, however, under certain conditions the change in

property rights due to the environmental legislation may make a farm more proﬁtable if

 

31 Similarly, a utility-maximizing Farm A could be at a point directly north of A. without the standard.
Again, the difference in proﬁts is an opportunity cost that Farm A bears due to the total maximum daily
load standard.

56

previously unexploited technologies that result in enhanced innovation offsets can be
identiﬁed and adopted. Indeed, despite the assumption of a trade-off between proﬁts and
environmental quality, induced innovations may be possible under certain conditions.
Whether innovation offsets and enhanced innovation offsets will be common and large,
or will be negligible, is an empirical question. The above simple analysis just clariﬁes
that they can exist when the ﬂexible ambient performance standard is strictly and

perfectly enforced.

When this last assumption is relaxed, so that now there is compliance uncertainty,
the resulting transaction costs may affect the possibility of induced innovation and
innovation offsets. Bounded rationality and imperfect information may inhibit farmers
from identifying a suitable technology that puts them on a production possibility frontier
like PPF2. Furthermore, compliance uncertainty prevents regulators from perfectly
enforcing a total maximum daily load standard. Thus, the incentive for farmers to meet
Wp, let alone seek a technology that will result in induced innovation, is removed. The
transaction costs resulting from compliance uncertainty due to asymmetric information,
moral hazard, imperfect information, and free riders may negate the potential beneﬁts of
ﬂexible ambient environmental performance standards, namely induced innovation and

innovation offsets.32

An alternative scenario would be the situation in which the regulator provides
farmers with a list of approved practices, or a technical sheet to meet the ﬂexible ambient
performance standard in an attempt to minimize transaction costs resulting from

compliance uncertainty. This technical sheet contains a list of practices that the regulator

 

32 The effects of strict enforcement of a high penalty can provide the incentives to innovate. This
possibility is examined in Chapter 6.

57

approves for use by dairy farmers. Farmers choose those practices that are best suited for
their farm situation. The regulator may believe that the technical list assures that the
speciﬁed water quality standard, Wp, will be reached, without too much concern for the

compliance costs incurred by farmers.

If the regulator does provide a technical sheet to farmers, it has converted the total
maximum daily load process from a ﬂexible ambient environmental performance
standard to a design standard. The manner in which the regulator implements the ﬂexible
environmental performance standard partially depends upon how signiﬁcant the
transaction costs resulting from compliance uncertainty are, and the regulator’s objectives

for improving water quality.

How will the regulator implement total maximum daily loads and how will
farmers react when they must comply with a ﬂexible ambient performance standard but
face compliance uncertainty? With informational constraints, both agents have the
incentive to behave strategically. The following chapter analyzes their actions within a
game-theoretical framework to highlight the possible outcomes that may arise, given their

different motivations.

58

CHAPTER 5

A Game-Theoretic Analysis of Farmers’ Choices

Having examined in Chapters 2 through 4 the factors important to understanding
the relationship between ﬂexible environmental performance standards and compliance
uncertainty, this chapter outlines the possible actions of the farmer who must contend
with a ﬂexible ambient environmental performance standard in the face of compliance
uncertainty. The analytical framework developed in this chapter reveals the effects of
uncertainty and its resulting transaction costs on the farmer’s compliance and innovation

choices.

5.1 A Game Theoretical Framework

Game theory attempts to study decision making where agents behave
strategically; it thus lends itself well to the analysis of compliance uncertainty within the
context of implementing total maximum daily loads as an environmental performance
standard. In the real world, some economic agents like regulators and farmers make
strategic decisions with respect to their actions. Strategic decisions are made because
both regulators and farmers have imperfect information concerning the consequences of

alternative actions.

A game can be deﬁned to represent a competitive situation where two or more
players pursue their own interests. Economic games are composed of: the players (or
economic agents), the rules of the game, the payoffs of the game, and the information
conditions that exist for the duration of the game.33 Speciﬁcally, the player is a decision-

making unit and can be an individual or an organization. The rules of the game describe

 

33 Adapted from Mansﬁeld 1988.

59

how the resources can be used. A strategy is a complete speciﬁcation of what a player
will do under each contingency of the game. Players receive a payoff which speciﬁes the
game’s outcome for the players. Each player’s payoff depends upon the strategies that

each player chooses to follow.

A simple game can illustrate the effect of compliance uncertainty and its resulting
transaction costs on the actions of Michigan dairy farmers -— whose farms are examples of
potential sources of nonpoint pollution '— in complying with a total maximum daily load

limit that is implemented as a ﬂexible environmental performance standard.

5.1.1 The’Players and the Rules of the Game

There’are two players in the game, the state regulator and one farmer, farmer i.
The resource of interest is a degraded surface body of water. The TMDL is the rule that
describes how this resource can be used. In order to illustrate the consequences of
compliance uncertainty, the TMDL will alternatively be implemented as an ambient
performance standard, an efﬂuent performance standard, and a design standard to control
pollution from nonpoint sources. Once a total maximum daily load process is
implemented, the level of pollutant can not exceed that speciﬁed by the TMDL standard.
As long as the TMDL is not speciﬁed as a design standard — that is, it is always a ﬂexible
performance standard — those owning nonpoint sources of pollution can choose any

method they wish to meet the standard.

For the purposes of illustration, in the following extensive-form game, the total
maximum daily load limit is taken as given; in other words, the optimal level of the
standard is not derived, and thus is exogenous to the model as are the economic beneﬁts

of pollution control. Note that the allowable level of pollutant speciﬁed by the TMDL

6O

and the associated beneﬁts may not be efﬁcient. As a consequence, economic efﬁciency
— the maximization of net social beneﬁts — is not the objective of this model. Rather, the
regulator and farmer are in a second best world, where the goal is to achieve the TMDL

standard cost-effectively.

In order to simplify the discussion, the following game assumes that the TMDL is
implemented as a ﬂexible ambient performance standard. The regulator only monitors the
ambient level of phosphorus in the degraded water body, and if it exceeds the allowed
limit, then all farmers who in the watershed will be penalized. In this illustration, the
farmer’s choice is depicted as a binary choice — comply or .do not comply — when in
actuality each farmer chooses an abatement level from a continuum of possibilities. This

assumption is re-speciﬁed in the next chapter.

5.1.2 The Expected Payoffs of the Game

The expected payoffs are incentives for the players. Ex ante, each player
considers the expected payoff because there are probabilities associated with the actual
outcomes. The payoff to the farmer is measured in dollar values and is actually
considered to be a cost. The farmer’s objective function is assumed to be the
minimization of his cost of his abatement effort plus an expected penalty that may be
assessed if the ambient pollution limit is exceeded and if he does not make any effort to
abate his efﬂuent. The regulator takes the social welfare point of view which means she34
values the resource owners’ interests (the farmer) and environmental concerns. As such,
the regulator is assumed to minimize her payoff (cost) which is the farmer’s expected

costs, subject to an environmental constraint that is deﬁned by the TMDL standard.

 

34 For notational simplicity, the regulator is assumed to be female.

61

5.2 Sequential Extensive-Form Games

A sequential extensive-form game is used in this ex ante analysis, similar to the
analysis used by others such as Segerson (1999a, 1999b) and McCluskey (2000). The
purpose is to visually illustrate the how compliance uncertainty affects the strategic
behavior of the regulator and the farmer when they must contend with a total maximum
daily load. Segerson uses this framework to analyze mandatory ﬂexible incentives to
control agricultural pollution (1999a), as well as comparing mandatory and voluntary
approaches to food safety. McCluskey uses a similar game-theoretic framework to
examine asymmetric information and policy issues relating to organic foods. In this type

of game, each player moves in turn, they do not take actions at the same time.

An extensive-form game is simply a decision tree representation of the sequential
game. This visual depiction makes it easier to envision the sequence of actions and
events, and will aid in understanding the more complex model to be presented in Chapter
6. The extensive-form game in Figure 5 stylistically depicts the possible outcomes when
a TMDL is implemented as a ﬂexible ambient performance standard for the special case

of a single farm.

The process of TMDL regulation is modeled as a four-stage game. First, the
regulator implements a TMDL process by choosing an ambient limit, a , as well as a
penalty, P , in case the limit is not met. The regulator chooses the values of these
variables in anticipation of the actions of the farmer. The limit, 5 , requires the nonpoint
source to reduce its efﬂuent of phosphorus loadings into the surface water body. This
limit may actually differ from the target mandated by the TMDL. Why can these two -

the actual limit and the TMDL target — diverge? The TMDL requires, states to ensure that

62

the designated water quality standard is met. Yet, the regulator can make the limit more
stringent than the designated standard. For example, if the regulator has little information
regarding what farmers are doing, but must ensure that the TMDL-mandated standard is

met, the regulator can make the actual limit more stringent.

In the second stage, the farmer chooses whether to comply and innovate, comply
and continue to use his current technology, or he can choose to not to comply in order to
minimize his compliance costs and the expected penalty.35 Why is this latter choice
possible? As Heyes (1998) states, since there is a cost associated with compliance,
whether a pecuniary cost or a transaction cost or both, regulations have to be enforced;
here, the environmental regulation has now increased the cost or price of water quality
from zero for the nonpoint sources. That is, water is no longer an unpriced part of the
farmer’s production process. But since monitoring and enforcement are costly, they are
usually incomplete. The farmer may take advantage of the situation and not comply if the
expected gains from non-compliance are greater than the expected penalty. Consequently,
some farmers are able to avoid compliance and may decide to free ride on the efforts of

those farmers who do try to meet the TMDL requirements.

 

35 Conceptually, the farmer chooses his degree of compliance to minimize his abatement costs plus the
expected penalty. The next chapter models this situation more appropriately.

63

KEAGQ .4... .~ Erick RNQQQ Q. .~ SEER \xeAeQ m. .a SEEK .mxobi .4. .~ $53k

 
  

    

   
 

 

.NBASQ .4183:me REGAN a. SBBmox @osteo. 9.183332 .QSASQ “.183:me
8:28; monzacom cam—econ 8&3:ko
8: meow pen 8: moo—o was
£0582 $0252 $8602 29:52

Sue—awom noun—awom
>
0:385
use
banU
meoo
8: on
58.8%
m . 1N1» 4932.
880:0 955295
53233“

  

KNQQQ .0. .~ .162th NFASQ .4. .n SEEK

§9€Q “.185:me §9€Q .4. KQBEMmz

 
  

       

ouzmcom
Ho: meow
maozcoE

mom=mcom
was
39:52

Susiwom

bcﬁtooca oocm=ano 53> 405:. m ”.0 onooSO new 2266. maimed 6 9:9“.

64

Thus, as shown in Figure 5, the farmer can choose to not comply due to the
regulator’s inability to perfectly enforce compliance. If the farmer chooses to not comply,
he does not incur any compliance costs. If the farmer does comply with the TMDL
standard, his costs increase, in other words his proﬁts decline (P1 to P2), as shown in

Figure 3, Profits and Environmental Quality Frontier, which is reproduced here for

 

 

convenience.
Figure 3: Proﬁts and Environmental Quality Frontier
Proﬁt
P1
P2

 

PPF,

 

 

 

 

4L
7

 

0 W1 We Environmental Quality

 

 

 

Adapted from Srivastava, Batie, and Norris (1999).

What are the incentives to comply and innovate in the face of compliance
uncertainty? The farmer’s is motivated to innovate in order to reduce his effort costs and
his expected penalty, that is, reduce his total expected compliance costs. If he does
innovate, then the Porter Hypothesis would in fact hold. Presumably the farmer chooses
to innovate because he has estimated that his compliance costs are minimized by doing

so. In this case, innovation offsets (at A4) and perhaps even enhanced innovation offsets

65

 

(at A5) may be realized by the farmer as depicted in Figure 4, Induced Innovation and

Innovation Oﬂsets, which is also reproduced here for convenience.

Figure 4: Induced Innovation and Innovation Offsets

 

A

Proﬁt

 
 
  

Enhanced Innovation Offsets

 

 

 

 

 

 

0 W1 W9 Environmental Quality

 

 

 

Adapted from Srivastava, Batie, and Norris (1999).

In the third stage, the regulator monitors the degraded water body to determine
whether the ambient level of phosphorus exceeds the mandated level, Zr“ . The regulator
experiences compliance uncertainty due to (a) unobservable efﬂuent, (b) incomplete
knowledge on their behalf with respect to understanding the effect of management
practices and technologies on the generation of efﬂuent and the fate and transport of
efﬂuent from nonpoint sources, (c) the actual fate and transport of efﬂuent and its
stochastic characteristics, and ((1) because farmers can behave opportunistically due to
asymmetric information between many actors and the regulator, resulting in moral

hazard.

66

Since the regulator is only monitoring the ambient level of phosphorus in the
water body, she does not actually know what each farmer is doing. Consequently, the
regulator can mistakenly penalize a farmer who has undertaken abatement efforts. Thus,
in the third stage, the regulator can monitor the ambient level and penalize the farmer or

monitor the ambient level but not penalize the farmer.36

In the fourth stage, the payoffs are received. The top row is payoff to the
regulator, and the second row is the payoff to the farmer. These payoffs are expected
payoffs since the there are probabilities associated with each branch of the tree, so the
payoffs are the products of the probabilities and the dollar amount. The desire to
minimize his costs —- that is, his expected payoff — is the incentive which motivates the

farmer in his decisions.

Although not depicted in Figure 5, the effect of lawsuits from third-party citizens’
groups can be an important factor on the farmer’s decision. Third-party citizens’ groups
have been the catalyst in ensuring the implementation of total maximum daily loads by
states and the EPA. Also, some third-party groups have used these lawsuits at the local
level to motivate both regulators and farmers in complying with the requirements of the
Clean Water Act. In Michigan, the Sierra Club has given lawsuit notices to Michigan
dairy farmers, and this appears to have increased concerns amongst dairy farmers
regarding their compliance decisions with environmental regulations (Jackson 2001).
Even if farmers would like to comply, liability issues may prevent them from investing in
the necessary research and development that may result in induced innovation and

innovation offsets. That is, compliance certainty is costly to determine with an ambient

 

36 If the regulator did not enforce the limit, then that would be an additional choice - to not monitor and
not enforce. But since it assumed here that the regulator will enforce the limit, this option is not speciﬁed.

67

standard; for example, where free riding is a concern. This situation compounds the
difﬁculty of knowing who has undertaken efforts in good faith. In this case, farmers may

face liability issues that are not due to their own efforts.

In such a situation, they may prefer to not undertake the necessary research and
development investments (i.e. they may not innovate), but would prefer to shift liability
costs to the regulator by demanding design standards speciﬁed by the regulator. Then by
following the design standards, farmers are still able to reduce their compliance costs, but
without having to be innovative. Since the regulator wants to minimize the costs that
farmers incur while ensuring that water quality is improved, the regulator may thus also
be motivated to convert the ﬂexible environmental performance standard into a design
standard. If this is the case, then there will be no induced innovation, no innovation
offsets, and the Porter Hypothesis will not in fact hold in the case where regulators try to
control nonpoint source pollution because of compliance uncertainty. In essence, liability
concerns due to compliance uncertainty could cause both the regulator and the farmer to
search for greater certainty, and want to move from a point like A to B in Figure 2,

Flexibility and Compliance Uncertainty Frontier, duplicated below.

The equilibrium concept used in this analysis is Nash equilibrium. A Nash
equilibrium exists if each player is employing his optimal strategy, given all the other
player’s strategies. With a Nash equilibrium, no player has an incentive to deviate from
its strategy, so the equilibrium is stable. Thus, with a Nash equilibrium, each player is
satisﬁed that he has made the best decision possible for himself or herself, given the
decision of other players, and has no incentive to change his or her decision. In deciding

which course of action to take, the farmer compares his alternative payoffs, and chooses

68

that one which allows him to minimize his costs, given his information set regarding the

regulator’s actions and the actions of other farmers, and the third-party citizen’s group.

Figure 2: Flexibility and Compliance Uncertainty Frontier

 

Flexibility “ Flexible Ambient Environmental Performance
Standard
A
Fl

   
 
  
 
 
   

Flexible Efﬂuent
Environmental Performance
Standard

Design Standard-based
Environmental Regulation

 

 

 

 

 

 

Compliance Certainty

 

Although Figure 5 outlines the possible courses of action that the farmer can take,
as is, it does not conclusively distinguish what choices the farmer will in fact make. The
course of action that the farmer chooses depends upon the likelihood he places on the
different events occurring. Additional factors that affect his subjective probability are the
degree of compliance uncertainty, the compliance measure, the regulator’s monitoring
and enforcement efforts, the expected penalty, and any expected legal costs. Ex ante, the
farmer also forms a subjective probability that the actual level of phosphorus in the water
body exceeds the limit, (7 , which in turn is a function of the actions of the farmers

around the water body and stochastic environmental drivers.

69

This analysis is not intended to be a comprehensive reﬂection of all the intricacies
in this situation; rather, it is intended to provide insight into the incentives that the farmer
faces when ﬂexible environmental performance standards are used to control nonpoint
source pollution. In reality, farmers are not likely to make a dichotomous decision of
complying or not complying. Rather they are more likely to undertake some level of
abatement effort that will allow them to be in compliance at least cost. The next chapter —
Chapter 6 - investigates this more realistic representation. Ultimately, the farmer’s effort,
his innovation decision, and whether the Porter Hypothesis holds for the case of nonpoint
sources of pollution, depend upon the effects of compliance uncertainty and its resulting
transaction costs, as shown in the next chapter, Chapter 6, Environmental Standards and

Compliance Uncertainty.

7O

CHAPTER 6

Environmental Standards and Compliance Uncertainty

The model developed in this chapter allows an examination of the differing levels
of cost-savings due to innovation that result from various levels of compliance
uncertainty with a ﬂexible performance standard. Speciﬁcally, this model is used to
investigate the implementation of a TMDL process to control phosphorus from dairy
nonpoint sources in the presence of compliance uncertainty. The TMDL process is
implemented alternatively as an ambient standard, an efﬂuent standard, and a design

standard to investigate and compare the incentives under each approach.

As implied by Porter’s Hypothesis, ﬂexible environmental performance standards
may ensure that the desired environmental outcome is achieved while providing farmers
with the incentive to seek out the least-cost set of effective pollution prevention practices
for their unique farm situation. As discussed in Chapter 4, due to this induced innovation,
farmers may be able to decrease their compliance costs through innovation offsets
resulting in clear beneﬁts for farmers while also improving environmental quality. It is
unclear, however, whether these innovation offsets will be possible in the presence of

compliance uncertainty.

The six sections of this chapter are divided as follows: the ﬁrst section speciﬁes
the structure of the model, the second considers a base case of where there is no
compliance uncertainty. The third section turns to the more realistic case of imperfect
information and compliance uncertainty when a TMDL process is implemented as an
ambient performance standard. The outcome of an efﬂuent standard is investigated in the

fourth section, while the ﬁfth looks at the last case of a when the TMDL process is

71

 

 

implemented as a design standard. The ﬁnal section summarizes the ﬁndings from this
chapter, providing a ranking of the costs from these different compliance measures. In so
doing, Chapter 6 brings together the various factors and issues discussed in earlier

chapters.

6.1 The Structure of the Model

This section introduces assumptions and notation used in the model that is
analyzed. The focus throughout is on the effectof compliance uncertainty on the
implementation of environmental performance standards. Despite the apparent
advantages of ﬂexible environmental performance standards in terms allowing induced
innovation, the practical difﬁculties of this approach are highlighted once compliance

uncertainty and its resulting transaction costs are taken into account.

6.1.1 Phosphorus in the Lake

It is assumed that the water quality of a lake is degraded by efﬂuent by a single
pollutant — phosphorus. The only sources of water efﬂuent around the lake are nonpoint
sources, and are all assumed to be dairy farms. Assume there are numerous dairy farmers,

i = 1,2, ...,n, who each generate some quantity of phosphorus, r, , as a result of their milk
production. Phosphorus levels depend on a farmer’s abatement efforts, x,, on the current
technological state, 6,. , and stochastic weather events, vi; that is, r, = (x,,6,.,v,). Effort
costs are denoted c,(x,.,6,. ). Marginal costs increase when the farmer increases his efforts,
(8c, / 8x, > 0,\7’i). Farmers are price takers in input and output markets, and farmers in the

watershed are assumed to have no collective inﬂuence on input or output prices. Each

farmer is assumed to be a risk-neutral, cost-minimizer.

72

Following Horan, Shortle, and Abler (1998), the observed ambient concentration

of phosphorus is given by:
a = (1(rl ,..., r,...rn , g, xi)

where a is the ambient concentration, r,- (i = 1,2,. . .,n) is efﬂuent from nonpoint source i

(da / 6r, > 0, Vi), g represents stochastic environmental variables that affect the transport

and fate of phosphorus, and his a vector of watershed characteristics.37 Note that the
ambient concentration is stochastic in nature since it is a function of g and r,-, each of

which is stochastic.

The regulator’s objective is to ensure that the TMDL standard is met in order to
improve water quality. The regulator ensures compliance through monitoring and

enforcement efforts by levying a penalty on those who do not comply.

6.1.2 Information Conditions

Following Horan, Shortle, and Abler (2002), let (2 be the information set that
encompasses all relevant information about both random and non-random processes with
respect to pollution of the lake. Neither the regulator nor the farmer will have full
information; rather, each has access to only a subset of the information in Q. Although,
the regulator is unable to perfectly observe r,-, or the efﬂuent, from the nonpoint sources
at a reasonable cost, it can form expectations which are conditional on observations of the

farmers’ management practices and past weather information.38 Denote the regulator’s

 

37 Phosphorus pollution is a dynamic process, but it is modeled here as a static process for simplicity.
38 These expectations may be assisted by bio-physical models of nonpoint source pollution.

73

information set of the farmers’ actions and the effects of their management practices on

water quality as 0,.

Similarly, farmer i possesses an information set, 0, e 0, regarding his actions, the

efﬂuent process (both random and nonrandom components) and the effects of his
management practices on water quality. For an ambient standard, this information set also
pertains to the actions of others and their effect on ambient pollution levels. In general,
the farmer has better information about his management practices and the regulator has
better knowledge of the effects of these practices on the lake’s water quality. The result is

an information asymmetry between the two players:
(2, at 0, Vi
and further,
9, at Q}. Vi ¢ j
where j is another farmer. As these information sets are incomplete, both the regulator
and farmer will experience uncertainty regarding the exact speciﬁcation of r,-(-).
Additionally, since efﬂuent from each farm i is a function of g, which represents

stochastic environmental variables that affect the transport and fate of phosphorus, the

uncertainty with respect to r,-(-) is exacerbated.

6.2 Performance and Design Standards with Compliance Certainty

Initially, consider the case of compliance certainty to provide a benchmark to

which to compare the effects of compliance uncertainty. Everyone has perfect

74

information. All functional relations are known by all players, and there are no stochastic

variables. Only when these assumptions are relaxed does compliance uncertainty emerge.

With compliance certainty, the regulator can impose a standard and know for
certain whether each farmer complies. The farmer also has no uncertainty about the
regulator’s ability to enforce the standard, which she can easily do with a sufﬁciently
large penalty for non-compliance. Under these circumstances, the ambient standard, and
efﬂuent standard, and a design standard could easily be set such that each farmer
complies in a manner that achieves the TMDL cost-effectively. In other words, the
outcome of a cost-effective design standard under perfect information will be the same
regardless of the type of compliance measure; therefore the focus here is on the case of a

design standard for simplicity since it is equivalent to the other two compliance measures.

A design standard in this simple model restricts x, = x, , leaving the farmer with

no ﬂexibility in making effort choices. Farmer i’s associated effort costs are illustrated by
area I +1] in Figure 6 where MCI is his marginal cost of effort curve. If the farmer
innovates in a way that reduces his marginal cost of effort, the corresponding curve will
shift down, to MCZ; his costs are reduced to area II, for a cost savings equal to area I
relative to his original technology. Hence, the cost savings of area I represents the

farmer’s incentive to innovate when there is no compliance uncertainty.

This assumption of perfect information, however, is unrealistic. The regulator
does not have detailed information about each farmer’s management practices nor his
level of abatement effort; moral hazard is a problem in the real world. Additionally,
farmers do not know what each other are doing. In the next section this assumption of

perfect information is relaxed.

75

Figure 6: Effort and Innovation with Perfect Information

 

 

 

 

 

 

 

$
MCl
MC2
I
II
it
Effort Level
Higher Effort Lower Effort

 

6.3 An Ambient Performance Standard with Compliance Uncertainty

The previous case showed that in the case of perfect information, a sufﬁciently
high, certain penalty can provide farmers with enough incentive to innovate, leading to
innovation offsets. In reality, however, the regulator’s information set is not equal to that
of the farmer’s, and no farmer knows what others are doing — perfect information does

not exist — and in fact there is compliance uncertainty ( Q, :6 Q, and Q, ¢ Q}. ). Consider

the case where the regulator implements the TMDL process as a ﬂexible ambient
performance standard. The regulator only measures the total ambient concentration of
phosphorus in the lake, a , before making a decision on enforcement. But now with
imperfect information, the regulator can not identify which farmer contributed the most
or the least to the total ambient concentration due to the random variables involved and
incomplete information regarding the fate and transport process. Additionally, with

imperfect information and hence compliance uncertainty, no farmer is certain that his

76

abatement efforts are sufﬁcient to ensure that the ambient standard is met, and that he

will not be penalized.

The regulator implements the total maximum daily load process as a ﬂexible
ambient performance standard to control phosphorus. The regulator’s objective function
is:

Min Eerie-,9.) (1)

E,Penalty i=1
S.t. Pr(a >aTMDL)< z , 214(67,Penalty) Vi .

In Equation 1, 05mm , is the ambient phosphorus standard speciﬁed by the TMDL (the
TMDL target), 2 is a probability target between 0 and 1, and 5c," ((7 , Penalty) is the

regulator’s expectation of farmer i ’3 response to the policy variables. The environmental
constraint is speciﬁed probabilistically due to the stochastic nature of the efﬂuent and the
ambient water quality level; additionally, the farmer may not be able (nor willing) to
ensure that the ambient performance standard speciﬁed in the TMDL is met at every
point in time. The limit, ‘cY , requires the nonpoint source to reduce its effluent of
phosphorus loadings into the surface water body. The TMDL target may be different
from what the regulator chooses to require the farmers to meet, 5? ; for instance, 5? may

be more stringent than aTMDL to increase the likelihood that the TMDL target is met. The

regulator chooses the values of these variables in anticipation of the actions of the farmer.

The limit, bi , and the TMDL target can differ because the state regulator can
choose to set the limit to whatever level she wants; she just must ensure that the

designated water quality standard is met, as speciﬁed by the TMDL target. Thus, if the

77

regulator has little information regarding what farmers are doing, but must ensure that the
TMDL-mandated standard is met and also wants to compensate for stochastic weather
events, then she can choose to make the actual limit more stringent than the TMDL

standard.

The regulator’s optimization problem is also subject to the regulator’s
expectations of the farmers’ responses to the penalty and c7 . The optimization problem is
as it is because imperfect information forces her to form expectations of what the farmer

will do based on her information set, (2, ,resulting in the regulator pursuing a second-

best outcome. A ﬁrst-best outcome would minimize total costs subject to the
environmental constraint, but not subject to the expectations of the farmers’ responses
since the regulator would know what these are. The ﬁrst-best outcome, since it is based

on better information, would be least cost over all second-best outcomes.

In the case of a ﬂexible ambient performance standard, farmer i ’s objective

function is:

Min 6, (x1. ,61) + Pr(a > 67|xl , 9: ,x2 ,6, ,...,xn , 6’" ) x Pr(enf0rcement|a > £7) XPenalty.

F arrner i selects his level of effort so as to minimize his expected costs which is the sum
of his costs of effort and the expected penalty; it is an expected penalty because recall, a
penalty is only imposed if a _>_ E . The expected penalty in turn is the product of the
farmer’s subjective probability that the ambient level of phosphorus, a , exceeds the

threshold speciﬁed by the regulator, E , the farmer’s subjective probability that the

78

regulator will enforce the penalty, and the pecuniary penalty.39 The ambient level of
phosphorus and hence the expected penalty faced by the i ’th farmer is dependent upon

the actions and technologies used by all n dairy farmers.

For notational ease, consider the case of Farmer 1. Given the initial state of

technology, 01' , his objective is to choose effort x,, to minimize expected costs:
Ml 71 c1 (x, ,6: ) + Pr(a > Elxl , 6,1,x2 , t9, ,...,xn , 6n ) x Pr(enf0rcement|a > Zr") xPenalty.
II

The solution to F arrner 1’s problem depends upon Zr" , his technology state, the actions of

all the other farmers since these jointly inﬂuence a , and the penalty level:
x1(Zz_,t9,',x2,62 ,...,xn , 6,, ,Penalty). (2)

In a Nash equilibrium involving n farmers, the farmer’s equilibrium solution is
x1(c7,6ll' ,Penalty) because every farmer’s solution depends upon what every other farmer

is doing; this solution reﬂects this interdependence. Under an ambient standard, the
regulator only monitors the ambient level of phosphorus and does not know with
certainty the abatement effort that any individual farmer is undertaking in order to
comply with the mandated limit, (‘1' ; that is, she does not have enough information to

perfectly estimate Equation 2.

In this case, if the regulator is facing pressure to ensure that the TMDL is being
met, she may set the penalty very high to ensure effort levels are high enough to satisfy

the TMDL standard. This possibility is consistent with the ﬁndings of Cabe and Herriges

 

39 There is no subsidy nor variable tax based upon a deviation from the TMDL standard. Subsidies can be
used in combination with ﬁnes to ensure compliance efﬁciently, as shown by Xepapades (1991). Subsidies
are not considered here to better reﬂect the intentions of the state of Michigan, where the Department of
Environmental Quality is not considering any subsidy program to facilitate TMDL compliance.

79

(1992) who determined that, with an ambient performance standard, the penalty may
have to be set very high to ensure compliance due to informational constraints; that is,
Cabe and Herriges suggest a high penalty because each producer may not believe that his
choices inﬂuence pollution levels at the margin, or he believes that the effect of his
actions is small, much smaller than the actual effect.40 In essence, the regulator is
requiring Farmer l to over-comply by mandating a limit, if , that is more stringent than

“mm. (c? <aTMDL ). Farmer 1 does know that the regulator is setting a high penalty, but

he may not understand how the process works nor know the effects of his management
practices on the ambient pollution level. He therefore forms expectations with respect to
how conscientiously compliance will be enforced by the regulator. These expectations
result in a steep expected marginal penalty curve denoted EMPAmbient in Figure 7.41 The
EMPAmbm curve is not vertical because of compliance uncertainty; that is, since there is
no perfect information, EMPAmbient is Farmer 1’s subjective marginal penalty curve. He
knows the regulator cannot perfectly enforce the ﬂexible ambient performance standard
so he forms a subjective probability of enforcement. Additionally, without perfect
information, Farmer 1 does not know the actions of the other farmers, so he can only
form a subjective probability of whether the ambient level of phosphorus will exceed the

allowed limit; these uncertainties are reﬂected in the non-vertical slope of the curve.

Farmer 1’s initial marginal cost of effort is given by the curve MC]. With a high

degree of uncertainty, the amount of abatement effort Farmer 1 expends is x” . The cost

 

40 The literature generally views ambient pollution-based incentives as providing poor motivations for
pollution control (Cabe and Herriges 1992, Horan et al. 1998, 2002). Additionally, farmers must process a
large amount of information in order to correctly evaluate incentives that ambient-based policies attempt to
provide.

41 A high penalty is not unrealistic; the Clean Water Act speciﬁes a maximum ﬁne of $25,000 a day for
every violation.

80

that he incurs for this level of effort is equal to the area x’bc, his expected penalty cost is
equal to the area abx", for a total expected cost equal to area abc."'2 If Farmer 1 were to

innovate, his objective function would now given by:

' c (x ,62)+Pr(a > (7x ,62,x ,6 ,...,x ,6 )xPr(enf0rcementa > c7)xPenalty
In I 1 l I I 2 2 n n

where 6,2 represents his technological state to reduce phosphorus loadings.

Figure 7: Effort and Innovation with a Flexible Ambient Performance Standard

 

 

 

 

$ EMPAmbicnt
MC1
1
b 1
MC2 l
d
e |
l
|
a I C
A A —
X 2 X x1 Effort Level
Higher Effort Lower Effort

 

 

 

Although Farmer l incurs some costs from searching, adopting, and learning the

new technology — his effort increases to x: — this innovation allows him to reduce his

 

42 It is possible that Farmer 1 will be sued if the TMDL standard is not met. He can form expectations that
his expected penalty — here the costs of the legal proceedings are included as part of his penalty — are so
high as to make the marginal expected penalty curve extremely steep and close to the Y-axis (or at least
rotated), where he could lose his farm. In order to avoid the possibility of losing his farm, he may take
precautions by expending considerable effort to reduce his abatement, thereby incurring very high costs. If
he can prove that he has met the limit, he may be able to shift the liability costs to the regulator. But
proving that he met the limit would be difﬁcult in the case of a ﬂexible ambient standard due to the effect
of the other n-I farmers on the ambient pollution level.

81

 

marginal costs, shifting his marginal effort cost curve to MCZ. If the farmer were to
innovate in a way that did not affect his compliance uncertainty — in other words did not
affect his expected penalty — he would reduce his total expected cost by area bdc. This

cost savings from the innovation is Farmer 1’s incentive to innovate.

Farmer 1’s expected costs have declined relative to the situation prior to the

innovation because now the cost of his effort is only area x2” dc and his expected penalty
has also been reduced to area adxgl. Note that part of what Farmer 1 used to pay — in

terms of his expected penalty costs — is now instead part of his effort cost (area xfdex‘ ),

transforming this amount from an expected cost to a certain cost.

If the innovation reduced Farmer 1’s compliance uncertainty, then his expected
marginal penalty curve, EMPAmbic'“, would shift down. This shift would reduce costs
further, thereby increasing the incentives to innovate. With a ﬂexible ambient
performance standard, however, this shift is unlikely to happen, especially when n is
large. If n were large, then at the margin, any increase in Farmer 1’s effort would not
likely change his subjective probability with respect to the ambient level of phosphorus
exceeding the speciﬁed limit, nor his subjective probability of enforcement, nor the
penalty level; that is, it is unlikely that there would be any agency response. A response
from the regulator is unlikely because ambient pollution, and hence the likelihood of a
ﬁne, depends upon what the other n-I farmers are doing. Recall that an ambient-based
ﬁne such as this one is a group ﬁne. If there are many farmers, with no one farmer being
a major contributor to the pollution problem, then changes made by one farmer can be

expected to have a minimal effect on ambient pollution levels.

82

These incentives to innovate under compliance uncertainty are greater than those
arising under no compliance uncertainty due to the steep expected marginal penalty. The
curve is steep due to the high level of compliance uncertainty, which results in the farmer
being forced to expend too much effort. With perfect information, the stande set by the
regulator is perfectly enforceable since the regulator knows what each farmer is doing,
thus, there is no need for over-compliance, or no need for a steep expected marginal
penalty curve. Consequently, while the cost-savings of area I in Figure 6 is the incentive
that drives farmer i to innovate, he has neither the incentive nor the freedom to adjust his

effort level further because by restricting farmer i ’s effort level to 36,. , the standard in

Figure 6 is like an inelastic marginal penalty curve. Even after the innovation, farmer i ’s

effort level remains at 2,.

With compliance uncertainty, the situation is different since the incentives to

innovate differ. A comparison is given in Table 3. Initially the farmer is at a higher effort
level (x’l > El) compared to the perfect information case because of the high expected

marginal penalty.43 If Farmer 1 does not adjust his level of effort after the innovation and
remains at {1, then his cost of effort would decrease since his marginal costs have
declined (xlec < x‘bc, Table 3), but his expected penalty cost would remain at abx’4 as
before the innovation, resulting in total expected costs equal to area abec (Table 3). The
incentives to innovate are equal to the area bce (his cost-savings); this area bce is greater
than that for the case of perfect information or compliance certainty (area I, Figure 6)

because of the higher effort level.

 

43 The restricted level of effort from Figure 6, f, , is reproduced in Figure 7, and is labeled 7c] in reference

to Farmer 1.

83

Table 3: Relative Ranking of Firm Incentives With and Without Uncertainty

 

 

Pre-Innovation Post-Innovation Post-Innovation
without with Adjustment
Adjustment

Perfect Information
Cost of effort I H! 11 not applicable
Savings relative to pre- not applicable I not applicable
innovation
Compliance
Uncertainty
Cost of effort x/‘bc xlec x 2" dc
Expected penalty cost abx"1 abx4 adxg'
Total expected cost abc abec adc
Savings relative to pre- not applicable bce bcd
innovation
Savings due to effort not applicable not applicable bed
adjustment
Relative Ranking 3 2 1

 

Source: Figure 6, Figure 7

But there are additional incentives under compliance uncertainty. The marginal
penalty curve is no longer vertical, so it imparts different incentives to farmers;
speciﬁcally, now Farmer 1 is able to choose his effort level by equating his marginal cost
of effort to his expected marginal penalty. With a vertical marginal penalty curve, Farmer

1 could not choose his effort level; he was required to expend 7c, level of effort.

The expected marginal penalty is similar to an ambient pollution tax in that the

farmer compares his marginal beneﬁts and costs in choosing his effort level. Thus, in

84

Figure 7, Farmer 1 alters his level of effort after the innovation, and actually increases it
from x“ to x," . Although this increases his effort cost (xlec < xfdc , Table 3), as with a
tax policy, F arrner 1 can beneﬁt from an adjustment to his effort level after innovating
because this adjustment lowers his expected penalty costs (adxg’ < abxl, Table 3). By

being able to adjust his effort level, Farmer 1 is able to reduce his expected penalty costs,
capturing an additional cost savings equal to the area bed (Table 3), which is unavailable
to him under the case of perfect information when he can not adjust his effort level.
Consequently, since the incentives under compliance uncertainty — with adjustment in
effort — provide the greatest cost-savings to Farmer l and hence the greatest incentives to

innovate, it is ranked highest in Table 3.

Compliance uncertainty provides Farmer l with additional incentives — in that by
adjusting his level of effort — he is able to reduce his expected penalty costs. These
additional cost-savings and hence incentives do not arise under compliance certainty.
While this analysis differs from traditional ones that assume no compliance uncertainty, it
is analogous to the ﬁndings of Milliman and Prince (1989) who show that after an
innovation has occurred and after the individual ﬁrrn’s marginal cost curve changes to
reﬂect these changes in cost, the innovating ﬁrm re-equilibrates its emission level under a
tax or other economic incentive. But such re-equilibrating does not occur under a design
standard.44 Milliman and Prince, however, do not highlight the role of compliance

uncertainty as providing the incentive to ﬁrms to alter their level of effort.

 

44 Milliman and Prince use the term direct control for a design standard.

85

6.3.1 Special Cases of Interest

There are two special cases of interest to this model; speciﬁcally, there are two
comer solutions possible in this model, one where the farmer does not comply and the
second where he exits the industry. If the regulator chooses to not enforce the speciﬁed
limit, 5 , and Farmer 1 knows this lack of enforcement to be the case, then his subjective
probability of enforcement is zero, effectively removing the penalty. Then the expected
marginal penalty curve will coincide with the X-axis; Farmer 1’s abatement efforts will
be at point c. At point c, the farmer does not comply with the speciﬁed standard, and puts
no effort into reducing his phosphorus efﬂuent. In this situation, farmers do not comply
with a total maximum daily load requirement, since there is no incentive to do so, thereby

negating any possibility of induced innovation and innovation offsets.

The other possible comer solution is that F armer 1 could exit the industry. The
regulator could set the penalty so high as to lead to a corner solution where the ﬁrm will
cease to be proﬁtable and will exit the industry (Cabe and Herriges 1992). If there is a
great deal of compliance uncertainty — such as with lawsuits and ambient standards — the
expected penalty could coincide with the Y-axis, so that the only way Farmer 1 could
comply would be to be at the point where his marginal cost of effort curve, MC‘,
intersects the Y-axis, where his abatement level is complete, or in other words, he exits
from the market. But since the regulator wishes to minimize the costs of abatement effort
by all farms, and not simply to drive all dairy farms out of business, a lower marginal
expected penalty is preferred. This preference can be achieved by decreasing the level of
uncertainty, thereby eliminating the need to set an extremely high expected penalty and

the possibility of all ﬁrms exiting the industry.

86

6.4 An Efﬂuent Performance Standard with Compliance Uncertainty

The regulator will seek to decrease the amount of compliance uncertainty for
herself and for the farmer in order to encourage compliance at a lower cost, given the
farmer’s objective function (Equation 1). One possible way that Michigan can implement
the TMDL process and increase compliance certainty is by converting the ambient
standard into an efﬂuent performance standard for each farm — perhaps by using a
computer model.45 An efﬂuent performance standard refers to the estimated amount of
phosphorous allowed from farm i, which is estimated based upon the production and
pollution control choices made by the farmer. Such a standard is ﬂexible because the
farmer can choose any method he wants to ensure that he does not exceed the standard or
threshold. Although Michigan regulators have not converted the TMDL ambient
performance standard for the water body (if ) into an estimated efﬂuent performance

standard for farmer i ( 7,. ), this case is considered since it is being done in other

jurisdictions as a means to reduce informational and transaction costs. This standard is
still a ﬂexible environmental performance standard since the farmer is free to adopt any
technology or production method. The difference is now he is told how much pollutant he

can discharge from his farm operation.46

The conversion from an ambient standard to an efﬂuent performance standard has

been done for other environmental regulatory objectives with varying degrees of

 

45 Some example computer models include EPIC, SWAT, AGNPS. All three models have been deve10ped
by USDA’s Agricultural Research Service. For further information, see http://www.ars.usda.gov/is/AR.

46 With an efﬂuent standard, it is possible that Farmer 1 is sued if the TMDL is not met. The effect of the
lawsuit would be to shift or at least rotate the marginal penalty curve up, so that the farmer has more
incentives to choose a greater level of effort in order to ensure that he does not lose his farm. But if he can
prove that he has met his efﬂuent standard, he can shift the liability costs to the regulator.

87

complexity elsewhere, such as La Crosse County, Wisconsin47 and in the Netherlands
(Ondersteijn et al. 2001), in an effort to counter some of the transaction costs associated
with nonpoint sources of pollution. The-La Crosse County Department of Land
Conservation has adapted a computer model developed by the University of Minnesota
that estimates the amount of phosphorus that leaves feedlots through runoff. This model
was adopted by the county to reduce monitoring costs. The Netherlands has adopted the
Mineral Accounting System (MINAS) which tracks nitrogen and phosphorus ﬂows on
individual farms and individually taxes every farmer for any generated nutrient surpluses
using a “farm gate balance approach.” It calculates the difference between nutrients

entering the farm (e. g. feed and fertilizers) and leaving the farm (Ondersteijn et al. 2001).

In developing a model that estimates phosphorus efﬂuent from farms, the
regulator must collect data — characteristics of the farm operation, bio-physical site
speciﬁc data, management information - and develop a model that incorporates its
assumptions. This ex ante informational transaction cost can be signiﬁcant, but this may

result in decreased compliance uncertainty for both the regulator and the farmer.

The effluent performance standard is differentially applied to all farmers; thus, the
efﬂuent performance standard for farmer i is not necessarily the same as that for farmer j .
It is assumed that the regulator divides the total standard amongst all farms in a manner
that accounts for herd size, the land area on which they apply manure, and various site
speciﬁc bio-physical characteristics, thereby improving the political acceptability of the

TMDL by farmers.

 

47 Personal communication with Chuck Zauner (Department of Land Conservation, La Crosse County,
Wisconsin) June, 14, 2000).

88

If the modeling was perfect and if the regulator shared this information with all
farmers, then all parties — the regulator and all the dairy farmers — would have perfect
information. In effect, if all farmers acquire the model and learn how to use it, each can
get immediate feedback about the effects of his management practices on environmental
outcomes; as a result, every farmer can make any necessary modiﬁcations and
investments to ensure that he meets the efﬂuent performance standard. Additionally,
since it is the same model used by the regulator, the farmer is able to perfectly anticipate
any forthcoming problems and take necessary actions to mitigate them. The results of this
case are analogous to those discussed in 6.2, with no compliance uncertainty, and so are

not investigated further.

But in reality, the modeling is necessarily imperfect, since for example, the
information about the exact speciﬁcation of the efﬂuent from farm i, r,-(') and the realized
weather are not precisely known. This information limitation results in some degree of
compliance uncertainty for both the regulator and the farmer. Additionally, sharing its
information set and model with farmers entails costs for the regulator — for example by
providing the computer software and training — as well as for the farmer. Furthermore,

the farmer may not trust the model, so Q, st 0,.

Consequently, there are still at least three sources of compliance uncertainty. Let
the efﬂuent standard for each farm be estimated from the model. The ﬁrst source of
uncertainty stems from the weather. Ex ante, before the weather is realized, the farmer
can use the model to learn the predicted (and allowed) efﬂuent from his operation, given
the regulator’s understanding of his practices, site characteristics, the forecasted weather;

and he makes his compliance decisions. Ex post, the weather is realized, and the regulator

89

enters this information into the model, and the regulator determines if the farmer has been
compliant. Thus, since compliance is determined after the farmer has made his decisions
and once the weather is realized, he never knows whether he really is in compliance with
the efﬂuent standard. Essentially, despite the efﬂuent standard, the farmer does not have a
meter at the end of his ﬁelds that accurately tells him what his efﬂuent is so as to be able
to adjust his practices as necessary. It is this lack of certainty that makes him unsure of

whether his abatement efforts are sufﬁcient for him to be in compliance.

The second source of compliance uncertainty is the inability of the regulator to
observe all the farmers’ choices. Moral hazard forces the regulator to either trust or
distrust what the farmer tells her that he is doing. If the regulator does not believe what
the farmer reports that he is doing, she can audit him, or auditing can be done randomly

to a selection of farms.

The third source of compliance uncertainty is the difﬁculty both the regulator and
the farmer can experience in trying to understanding the model. This cost may be greater
for the farmer, especially if he does not trust the model. Without intimate knowledge or
understanding of how the regulator models his farm’s operation, compliance uncertainty
exists for the farmer. Nevertheless, this cost may decrease over time as both agents learn

and understand how the model estimates and sets efﬂuent standards.

Now that the regulator has a tool — the model — to assist in estimating the
phosphorus efﬂuent from each farmer’s operation, she is able to assess his abatement
efforts in complying with the T MDL process. With a ﬂexible efﬂuent performance

standard, the regulator’s objective function is now:

90

_Min ice-,6.)

r.,Penalty i=1
1

S.t. Pr(a > aTMDL) < z , xf(i‘;,Penalty) Vi .

The second constraint, 2,5 (7;,Penalty) is the regulator’s expectations of the farmers’

responses to the policy variables, their individually prescribed ﬂexible efﬂuent

performance standard and the penalty.

Faced with an efﬂuent performance standard, Farmer 1’s objective function is:

Min C I (x1 , 9.1)4' PTO”, > 7, ) X Pr(enf0rcement|xl ) xPenalty 43
a}

where F, is the ﬂexible efﬂuent performance standard speciﬁed by the regulator for

Farmer 1’s farm. It is still an expected penalty because of the stochastic environmental
variables, and also since neither the regulator nor Farmer l have perfect information
about each other’s actions. The solution to Farmer 1’s problem now is a function of the

efﬂuent performance standard, the technology he is using, and the penalty:
x. (adrenalin. (3)

The regulator makes an informed guess as to what Farmer 1’s solution is
(Equation 3) using the computer model. The investments needed to develop estimates of
the efﬂuent from Farmer 1’s farm and the increased monitoring that must occur under an
efﬂuent-based standard result in less uncertainty for the regulator than would occur under

an ambient-based system. With decreased uncertainty, the need for a very high penalty is

 

48 The probability of enforcement only depends upon the farmer’s individual actions. Ambient pollution
levels could also inﬂuence this subjective probability, but it has a secondary effect, and for simplicity the
focus is only on the primary effect.

91

reduced. Rather, the regulator can set a lower penalty, accordingly represented by

13M?Efﬂuent in Figure 8.49

Figure 8: Effort and Innovation with a Flexible Efﬂuent Performance

 

 

 

 

 

 

 

Standard
$ .
EWAmbrent
MC1 / EMPEfﬂuent
/
b/
MC2 \f
d/
h <
/ > k
/
/
/
c
XA x82 XE Effort Level
Higher Effort Lower Effort

 

 

 

Farmer 1’s abatement effort is given by x5 (which is less than x4 in Figure 7), his
cost of effort is the area foc, his expected penalty is area afxE, for a total expected cost

equal to the area afc. If Farmer 1 were to innovate, his objective function becomes:

Min c, (x, ,012 ) + Pr(rl >71)>< Pr(enf0rcementla > b?) xPenalty

xi

where 612 represents his new technological state to reduce phosphorus loadings. With a

ﬂexible efﬂuent performance standard, his solution now depends upon his speciﬁed

efﬂuent standard, his new state of technology, and the penalty level:

 

49 EMPAlnbicm is shown as the dashed line in Figure 8.

92

x1 (ii ’612 9 Penalty) '

Assuming this innovation affects neither his subjective probabilities nor the

penalty level, this innovation causes his marginal cost curve to shift down to MCZ, and
results in a cost saving equal to the area hfc, despite an increase in effort level to x25 .
Farmer 1’s total expected costs have declined because his effort costs equal area

xfhc and his expected penalty costs have decreased to ahxf . The cost equal to area

xf hfxE is transferred from what was previously an expected penalty to a cost of effort; as

would be expected, the higher effort level translates into a lower expected penalty. With
the ﬂexible efﬂuent performance standard, since Farmer 1 wants to minimize his costs,

he will choose to innovate.

With an efﬂuent standard, it is more likely that the innovation may affect Farmer
1’s compliance uncertainty. If the technology were to cause efﬂuent to decline for a given
level of effort, and Farmer l believes that the regulator will be able to determine this
change through modeling, then Farmer 1 would expect that the probability of being
penalized will decrease. In other words, Farmer 1 would expect an agency response
(Milliman and Prince 1989), causing his expected marginal penalty curve to shift down

after the innovation to EMPZ, as illustrated in Figure 9.50

Now Farmer 1 actually decreases his level of effort relative to his effort level after
the innovation (sz ) to x3,51 reducing his total expected costs to area amc.52 As would be

expected, this lower level of effort translates into a lower cost of effort, area x 3mc, and

 

50 The original marginal cost of effort curve, MC 1, has been removed to simplify the diagram.

51 In actuality, x3 can be greater than or less than x5 since it depends upon the degree to which the expected
marginal penalty curve shifts; the shift illustrated is one possibility.

52 This is analogous to the agency response in Milliman and Prince (1989).

93

since the marginal penalty curve has shifted down, his expected penalty cost has also
decreased, and now is equal to area amx3. His expected cost savings relative to the case
where this shift does not occur — in other words his additional incentive to innovate due
to the shift — is equal to the area ahm. This additional incentive to innovate — area ahm —
can be termed the uncertainty-reduction eﬂect. This uncertainty-reduction effect

essentially results in lower effort and lower associated expected costs for the farmer.

Figure 9: Effort and Costs with Uncertainty-Reduction Effect

 

 

 

 

 

 

 

$
MCl EMPEfﬂuent
MC?- f EMP2
\>m
a c
E E

x 2 x x3 Effort Level

Higher Effort Lower Effort

 

 

 

As shown, these incentives to innovate under an efﬂuent standard differ from both
those arising under an ambient standard and the baseline case of perfect information. A
summary of these incentives is presented in Table 4. Compared to the case of perfect

information where the farmer cannot adjust his effort level, farmer i is initially at a higher

level of effort (x5 > it. ); but as with the ambient standard, the farmer has a greater

94

 

incentive to innovate because compliance uncertainty allows the farmer to adjust his level

of effort after the innovation.

Assuming no uncertainty-reduction effect — or that the expected marginal penalty
curve does not shift — Farmer 1’s incentives to innovate are equal to area fck (Figure 8) if
he does not adjust his level of effort. This cost-savings is greater than his incentives to
innovate under perfect information, area I in Figure 6, because of the higher level of
effort that is required with the efﬂuent standard due to compliance uncertainty. By

adjusting his effort level, Farmer 1’s total expected cost savings are equal to area fch.

Table 4: Comparison of Firm Incentives with Uncertainty-Reduction Effect

 

 

Expected Costs Ambient Efﬂuent Standard
Standard
Without With uncertainty-
uncertainty- reduction

reduction
Pre-Innovation abc afc afc
Post-Innovation adc ahc amc
Change in expected costs bcd fch afcm

(Incentive to innovate)

 

Source: Figure 6, Figure 7, Figure 8, Figure 9

With the uncertainty-reduction effect — or when the innovation is such that
compliance uncertainty is affected so that the expected marginal penalty curve shifts
down — the farmer’s incentive to innovate is even greater; it is equal to area afcm.
Regardless of whether there is an uncertainty-reduction effect, Farmer 1’s incentives to

innovate are greater with the efﬂuent standard than with the case of perfect information

95

 

because of compliance uncertainty which results in the need for over-compliance through

a steep expected marginal penalty curve.

In comparison to the ﬂexible ambient performance standard, there is greater
certainty, so F arrner 1’s expected penalty costs are lower, requiring him to initially
expend less effort with a ﬂexible efﬂuent performance standard (it/4 > xE). His initial total
expected costs are less (abc > afc, Table 4) because his expected penalty costs are lower
by area abf (Figure 8) and his effort costs are less as well with the ﬂexible efﬂuent
performance standard (x’bc > xffc). His effort level is less because the expected marginal
penalty with the efﬂuent performance standard is not as high as with the ambient

performance standard since the regulator’s level of uncertainty is lower.

The lower expected marginal penalty curve implies that even if the innovation
causes the marginal cost curve to shift down in the same manner under both a ﬂexible
ambient and a ﬂexible efﬂuent performance standard, after the farmer re-equilibrates his
marginal costs and his expected marginal penalty, his expected penalty and new level of

effort after the innovation will be less under the ﬂexible efﬂuent standard, so adc > ahc

(Table 4) by area adh (Figure 8). Both his expected. penalty costs (ahxf) and the cost of

his effort (xfhc) are lower with the ﬂexible efﬂuent performance standard. Since his

total expected costs are less under a ﬂexible efﬂuent performance standard, the farmer
will prefer this compliance measure over a ﬂexible ambient performance standard.
Additionally, he will have greater incentives to innovate under both of these measures

than if there is perfect information and no compliance uncertainty.

96

But if the innovation causes the expected marginal penalty curve to shift down —
that is there is an agency response and an uncertainty-reduction effect — then the cost
savings to Farmer l are even greater because his expected penalty and hence effort level
have been further reduced. This lower effort level means both his effort costs and his
expected penalty costs are also lower (adc > amc, Table 4). This relationship holds even
if the expected marginal penalty curve shifts such that Farmer 1’s new level of effort X3 is

actually greater than XE.

Without a change in his expected marginal penalty curve, the ﬂexible ambient
performance standard provides more incentives to innovate — except for the special case
of a comer solution. But, since his total expected costs are lower with the ﬂexible efﬂuent
performance standard, the farmer will prefer this compliance measure over the ﬂexible
ambient standard — regardless of whether he believes his innovation results in an agency
response as suggested by Milliman and Prince (1989). Furthermore, with the ﬂexible
efﬂuent standard, if his innovation does decrease compliance uncertainty, then his cost-
savings are even larger, possibly resulting in greater incentives to innovate than with the
ﬂexible ambient standard. Farmer 1 will choose those options that allow his expected
marginal penalty curve to shift down, even if it results in X3 > xE , because X3 still will be
less than xf . Given the ambiguity of the agency’s response and movement of the
expected marginal penalty curve, a clear ranking of these ﬂexible compliance measures is
not possible. Nevertheless, Table 5 does provide a ranking of these compliance measures,
where 1 indicates the measure that provides the greatest incentive to innovate and 4 the
least. The incentives to innovate reﬂect the cost-savings of the measures. Thus, the

compliance measure that results in the greatest cost savings is ranked the highest. Since it

97

is unknown whether the measure will result in an uncertainty-reduction effect, it is not
clear whether the ﬂexible ambient performance standard can be ranked higher than the
ﬂexible efﬂuent performance standard; as such, the ambient and the efﬂuent standard
with the uncertainty-reduction effect are each ranked 1/2.

Table 5: Ranking of Compliance Measures

Perfect Ambient Efﬂuent Standard
Information Standard

 

Without 'With
uncertainty- uncertainty-
reduction reduction
Incentives to Innovate 4 1/2 3 1/2

 

Given Farmer 1’s objective to minimize his compliance costs and the regulator’s
as well to minimize these costs across all farmers, both agents will want to have a lower
marginal penalty curve. The following section examines a compliance measure that

enables both parties to get the certainty that they seek.

6.5 A Design Standard to Reduce Compliance Uncertainty

When regulating nonpoint sources of pollution, the regulator can decrease
compliance uncertainty by implementing the total maximum daily load process as a
design standard. Here, the regulator provides farmers with a list of approved practices, or
a technical sheet. This technical sheet contains a list of practices that the regulator

endorses for use by dairy farmers.53

With a design standard, the regulator’s objective function is now:

 

53 By using a design standard to implement the TMDL process, the regulator has implicitly made it a
negligence rule. With a negligence rule, the responsible agent is penalized only if it fails to comply with the
“due standard care” regarding the pollution generating activity (Segerson 1995). A farmer who uses the
technical sheet is deemed to have complied with due standard care.

98

Min :c,(x,,6,)

3?,Penalty i=1
S.t. Pr(a > am”) < z , if) (f, Penalty) Vi

where the regulator is requiring farmers to choose a state of technology from an approved
list, represented by a? . This approved technical sheet could be technologies that are

approved by an agency like the Natural Resource Conservation Service (NRCS). The
term 55,0 (7c, ,Penalty) is the regulator’s expectations of the farmers’ responses to the
policy variables: the farmer’s effort level, and the penalty. The expected response,

2,0 (i, , Penalty) , may differ from the standard if there is imperfect monitoring.

Since there is a list of approved technologies from which Farmer 1 can choose
what his best for him, he is able to choose his level of effort. Farmer 1’s objective

function is:

Ml )1 Cl (161,6l ) + x Pr(enforcement|x1)xPenalty 54
9|

where 6l is in the current state of technology. The solution to his problem with a design

standard becomes:
x1 (6, , Penalty) .

Now Farmer 1’s solution is only dependent upon the state of technology and the penalty
level. By equating his marginal costs to the expected marginal penalty, Farmer l chooses

his level of effort, XD.

 

54 Again, the focus is only on the primary effect — the farmer’s individual actions — for this subjective
probability.

99

 

With the technical sheet, the regulator removes much of her compliance
uncertainty since now she can easily determine who is in compliance by simply
ascertaining whether the farmer is employing practices speciﬁed on the technical sheet,
thereby reducing some ex post transaction costs. Nevertheless, there are still sources of
compliance uncertainty, from stochastic variables like weather. Additionally, monitoring
will necessarily be imperfect since it is costly (Heyes 1998); consequently, farmers may

be able to deviate from the mandated level of effort.

Figure 10: Effort and Costs with a Design Standard

 

EMPAmbient
I

lEMPEfﬂuent

EMPDt‘lS"

 

 

 

 

Effort Level

Higher Effort Lower Effort

 

 

 

Aside from the beneﬁt of minimizing compliance costs for farmers, the regulator
may prefer to convert the ﬂexible environmental performance standard —- both ambient
and efﬂuent performance standard — into a design standard, especially if she is interested
in process and believes that the technical sheet will ensure that water quality goals are

met. With this greater degree of compliance certainty, the regulator can set the policy

100

parameters, 3? (the technical sheet) and the penalty such that the expected marginal

penalty curve is not so steeply sloped, to EMPDeSig", illustrated in Figure 10.55

What of the farmer? With the design standard, the farmer can determine easily
whether he is in compliance with the total maximum daily load requirement since he is in
compliance as long as he is using some subset of the practices on the approved technical
sheet. In fact, the design standard with a technical sheet is the only compliance measure
that allows farmers to determine for themselves whether they are in compliance with the
TMDL process. Since the farmer does not incur any search costs when he is provided
with a technical sheet, these informational transaction costs are reduced; in other words,

his effort level is lower relative to the ﬂexible performance standards.

Farmer 1’s effort level is xD and his expected total costs are equal to area aoc. By
reducing some compliance uncertainty, a design standard decreases Farmer 1’s
compliance effort relative to the ﬂexible ambient and ﬂexible efﬂuent performance
standards (xA > xE > xD), thereby reducing his associated total expected compliance costs
(abc > afc > aoc). A technical sheet also reduces Farmer 1’s subjective probability with
respect to a penalty cost that includes legal costs from a lawsuit. Farmer 1 may face
lawsuits if the TMDL standard is not met. But note that in providing a technical sheet, the
regulator implicitly allows the farmer to shift liability costs to her, causing the farmer’s
marginal expected penalty curve to shift down towards the X-axis. The farmer can not be

sued by a third—party group, since the regulator has approved the practices; if the farmer

 

55 The expected marginal penalty curve is drawn thusly for illustrative purposes. The expected marginal
penalty curves for the ambient and efﬂuent standard are also included.

101

happens to be sued by a third-party group, the farmer simply refers the aggrieved party to

the regulator.

The increased compliance certainty provided by the technical sheet can be
extremely important for the farmer if he is concerned with legal liability. Moreover, for a
farmer who believes that liability concerns will likely result in him losing his farm, the
ability to shift liability costs to the regulator may be the foremost beneﬁt to him of the
technical sheet. This improvement in compliance certainty may therefore improve the

political acceptability of the TMDL process amongst farmers.

It is unknown whether innovations under ﬂexible performance standards will shift
Farmer 1’s marginal cost curve in such a manner that his post-innovation cost-savings are
greater than those that he enjoys under a design standard. Consequently, nothing can be
said as to whether Farmer 1’s expected total costs under a design standard will be less
than his costs after an innovation with a ﬂexible ambient or efﬂuent performance
standard. Nevertheless, it is clear that with the ﬂexible performance standards, the
incentives to innovate are greater than with this design standard. In fact, with this design
standard, there is no room for innovation. Although Farmer 1 may discover a more cost-
effective method to meet the standard or shift down or rotate his marginal penalty curve,
he would have to get the innovation approved and placed on the technical list, and this
requirement would take time, as is the current situation with NRCS-approved

technologies.

If the farmer is able to innovate in a way that affects his compliance uncertainty ——
that is, causes a downward shift to his expected marginal penalty curve — then the ﬂexible

Performance standard alternative that allows for this uncertainty-reduction effect may be

102

preferred to a design standard; Figure 9 shows that the incentives to innovate — or the

cost-savings - are substantial with the uncertainty-reduction effect.

6.6 Implications for the Porter Hypothesis

What of the Porter Hypothesis? Although some ﬂexibility is retained with a
technical sheet because farmers choose those practices that are best suited for their farm
situation, there is no longer room for innovation. With a design standard, the regulator
and farmer move from a point like A to B in Figure 2, Flexibility and Compliance
Uncertainty Frontier (reproduced here) removing innovation as an option; however, the
loss of the opportunity to innovate may not be as important as the need to reduce
compliance uncertainty. This conclusion is especially true when lawsuits are possible.
This ex ante analysis indicates that Farmer 1 may prefer compliance certainty over the
opportunity to be able to innovate with a ﬂexible performance standard — particularly
with the threat of lawsuits and high penalties for non-compliance. He may prefer a design
standard over a ﬂexible performance standard, regardless of whether it is an ambient or
an efﬂuent compliance measure. This desire to decrease compliance uncertainty renders
the Porter Hypothesis immaterial, innovations are precluded, as are innovation offsets.
These analytical ﬁndings are corroborated with empirical data in the next chapter,

Chapter 7, Interview Findings.

103

Figure 2: Flexibility and Compliance Uncertainty Frontier

 

 

Flexibility A

F1

   
 
   
  

Flexible Ambient Environmental Performance
Standard

A

Flexible Efﬂuent
Environmental Performance
Standard

Design Standard-based
Environmental Regulation

 

 

 

 

 

 

Compliance Certainty

 

104

 

CHAPTER 7

Interview Findings

The analytical outcomes in Chapter 6 were corroborated against interviews with
state regulators and focus groups with dairy farmers in Michigan. The game theoretical
framework developed in Chapter 5 was the foundation for the interview questions with
Michigan regulators and dairy farmer focus groups. The analysis in Chapter 6 was
corroborated with the results from these interviews. Both state regulators in Michigan and
dairy farmers were consulted in order to ascertain the consequences of uncertainty on
compliance with a TMDL process, and the effectiveness of ﬂexible environmental
performance standards to control nonpoint source pollution. The objective of these
interviews was to determine how reasonable these outcomes were when compared with

the concerns and goals of “real world” economic agents.

7.1 Data Collection from Regulators and Dairy Farmers

Informed interviews with representatives from the following agencies were
conducted to determine how total maximum daily loads were being implemented in
Michigan: Michigan Department of Environmental Quality, the Michigan Department of
Agriculture, and the Natural Resources Conservation Service. These interviews were

conducted from December 1999 until Spring 2000.

Data were collected through focus groups with dairy farmers in each of the six top
milk-producing counties in Michigan: Clinton, Sanilac, Huron, Allegan, Ottawa, and
Ionia. The participants were identiﬁed with the assistance of Michigan State University
Extension Dairy Agents. In total, 47 dairy farmers participated in these focus group

discussions, which ranged in size from 5 to 12 dairy farmers. While the selection process

105

was not random and the focus groups are not statistically representative of dairy farmers
in the state, the individuals interviewed had, for the most part, consistent conclusions and
therefore may represent the views of many others. The average number of milking cows
owned by the participating farmers was 163; the smallest herd was 15 head and the
largest was 1550. The state average herd size is approximately 81, while the greatest

number of milking cows are on farms in the 100-199 herd size range. (Census of

Agriculture 1997).

7.2 Results of Interviews with Regulators

A list of the issues that was discussed with regulators appears in Appendix A. The
topics discussed included such matters as: how will compliance be determined, how
loading reductions will be divided amongst dischargers, and how monitoring and
enforcement will be carried out. Across all the agencies, the deﬁnition of compliance was

highlighted as being the most problematic issue, providing some empirical evidence for

compliance uncertainty for regulators.

7 .2.1 Few Implementation Details

Regulators indicated that few regulatory details had been thought out, let alone
reached. They had not decided how the TMDL process would be implemented, whether it
would be as a ﬂexible ambient standard, a ﬂexible efﬂuent standard, or a design standard.
The regulatory agencies conﬁrmed that they have little or virtually no information on
what farmers were doing. Consequently, the cases in Sections 6.3 and 6.4 with

compliance uncertainty are reﬂective of the situation in Michigan, although the lack of

106

 

information on farmers’ practices suggests that a ﬂexible efﬂuent standard is unlikely to

be the compliance measure in the immediate future.

The interviewed regulators were vague and unsure regarding the deﬁnition of
compliance with respect to the total maximum daily loads process, regardless of the type
of compliance measure. On the one hand, they expressed signiﬁcant interest in the
potential ﬂexibility embodied within the process, since it would allow them to deal with
different situations and site speciﬁc factors as necessary. They recognized that “one size
will not ﬁt all” farm and environmental situations. Nevertheless, it appeared that the
regulators preferred to assess the practices — or the technologies — used to determine
compliance, as opposed to actually monitoring ambient water quality in degraded water
bodies as a means to reduce their transaction costs. This preference indicates that it is
unlikely that a ﬂexible ambient performance standard will be implemented, instead there

is a preference for a design standard.

Indeed, at least one regulator claimed that he does not want a “Cadillac Plan” to
monitor and enforce environmental performance standards; rather he wanted a technical
list that could be used to easily check whether a farmer is in compliance. This desire may
stem from the possible but faulty assumption held by the regulator that it is known which
technologies will ensure that the desired water quality outcomes are met; hence regulators
do not emphasize the need to check actual water quality outcomes. These remarks
suggest that interviewed regulators want to implement the total maximum daily load
process as a design standard, as examined in Section 6.5. In this situation, there would be
no innovation, although some compliance uncertainty transaction costs are reduced for

both the regulator and the farmer.

107

Regulators also felt that farmers want to know clearly what compliance means,
that they want a streamlined system of determination, and need technical help to achieve
compliance. A straightforward technical list of approved practices would fulﬁll these
requirements for the farmers, as it is the only compliance measure which farmers can
easily determine whether they are in compliance with the TMDL process. One regulator
commented that such an environmental process would be more predictable for farmers
than their neighbors. The implication was that an approved technical list would allow the
farmer to minimize his exposure to liability from third-party groups (such as irate
neighbors), and more easily plan expected business expenses. In addition, regulators feel
that by using approved practices, farmers will be able to indicate to other levels of
government that they are “sanctioned” by the state and are conducting their business as
they should; for instance, this “sanctioning” may help them to meet township zoning
requirements. A design standard increases certainty, lowering these types of expected
costs for the farmer, lending evidence that the expected marginal penalty does decrease
for the farmer, as discussed in Section 6.5, reducing the farmer’s effort level and his

associated costs, given his marginal cost of effort.

7.2.2 Minimizing Monitoring and Enforcement Costs

Regulators indicated a desire to minimize monitoring and enforcement costs.
Although the level of enforcement will be determined by the public, one regulator
mentioned that they did not want to be “policemen” looking to punish farmers. Indeed, it
had not been determined which agency would undertake monitoring activities and
various personnel indicated that the resources had not been allocated in their budgets; one

regulator stated that monitoring activities may be contracted out to private ﬁrms.

108

This desire to minimize monitoring and enforcement costs, coupled with virtually
no information on what farmers are doing and no technical sheet, corresponds to the case
of a ﬂexible ambient standard in Section 6.3 in Chapter 6. By setting a very high
marginal penalty, all the regulators must do is monitor the ambient level of phosphorus in
a degraded water body to check if the imposed standard is exceeded. In order to minimize
enforcement costs, the regulator can simply penalize every farmer who could potentially

discharge into the water body.

But this desire to minimize ex post monitoring and enforcement costs suggests
that regulators may prefer to use a design standard as the compliance measure, at least
until they invest sufficient resources in human capital and the necessary infrastructure to
implement a ﬂexible performance standard. In order to implement a design standard, they
could use the technical speciﬁcations available through the National Resource

Conservation Service.

7.2.3 Move Towards a Design Standard

In terms of innovation, regulators clearly indicated that little room is envisioned
for farmers to be innovative in terms of the technology that they can use to comply with
the ﬂexible environmental performance standard. Thus, these interviews suggests that
Michigan regulators are interested in moving to a design standard, or implementing a
TMDL process with an approved technical sheet, thereby removing a great deal of
ﬂexibility from the total maximum daily load process. Since innovation is not a priority
for regulators, the loss of ﬂexibility in the TMDL process is not a concern. Consequently,
Michigan regulators seem to prefer point B in Figure 2, Chapter 3 to point A on the

ﬂexibility and compliance trade-off frontier (the ﬁgure is reproduced below). Although

109

 

they want to decrease ﬂexibility in the process, the trade-off is that they increase the

degree of compliance certainty with the TMDL standard, and forgo innovation.

Figure 2: Flexibility and Compliance Uncertainty Frontier

 

Flexibility 0 Flexible Ambient Environmental Performance
Standard
A
Fl .

    
 
  
  
  

Flexible Efﬂuent
Environmental Performance
Standard

Design Standard-based
Environmental Regulation

 

 

 

 

 

 

Compliance Certainty

 

But this desire by regulators for a design standard may only be true for the short
run, when the TMDL process is ﬁrst implemented. The lack of institutional surety in the
compliance measure suggests that Michigan regulators are open as to how the TMDL
process may be implemented, and allows for the possibility of an adaptive policy
approach. The interviewed regulators felt that as they gain experience and knowledge in
dealing with nonpoint sources and are able to reduce transaction costs associated with
implementing the TMDL process, eventually a ﬂexible performance standard that relates
to water quality outcomes will be employed. In time, the necessary resources may be
allocated to estimate runoff from farms, thereby implementing the TMDL process as a
ﬂexible efﬂuent standard. Eventually, then, farmers may be able to beneﬁt from induced

innovation and innovation offsets.

llO

7.3 Results of Dairy Farmer Focus Groups

Dairy farmers were also interviewed through a series of six focus groups. Each
dairy focus group began with an explanation of what total maximum daily loads are, then
focused on questions related to water quality issues, including third-party lawsuits.
Participants were also asked to complete a short written survey, individually, after the

discussion. A copy of the focus group script and the written exit survey is in Appendix B.

The farmers in the focus groups expressed numerous concerns with regard to the
implementation of total maximum daily loads. These concerns ranged from how these
water quality regulations will affect their farm practices and proﬁts, to how these rules
will be implemented, as well as whether these efforts will actually result in improved
water quality. The details of these discussions were published in the Michigan Dairy
Review.56 Despite the wide range of concerns that were expressed across the focus group

discussions, a common theme emerged relating to uncertainty.57

7.3.1 Desire to Reduce Uncertainty

Participating farmers wanted to know What will compliance with total maximum
daily loads mean for my farm? As of now, the answer is not known because the TMDL
process is just beginning in Michigan. The participants mentioned aspects of the

regulatory process that they believe must be included to reduce uncertainty. These criteria

 

56 See http://wwwirec.msu.edu/agecon/gmith_endowment/documeng/Daiw.PDF or
http://www.m_su.edW~mdr/archives/mdwol6nol .pdf for a copy of the article.

57 It should be understood that a summary of the comments made is presented. Not all participants even
necessarily agreed with these comments, and some disagreed.

lll

are shown in Table 6, and include such things as clear water quality goals and the desire

for guidelines as to how to comply.58

As predicted, the desire for certainty was also identiﬁed by the participants when
asked why they would undertake abatement efforts with a TMDL process in the exit
survey (Appendix B). Table 5 shows the rankings of the top three motivations.
Participants indicated their desire to be responsible stakeholders as the main reason for
undertaking efforts out of ten possible reasons. Nevertheless, farmers rated the need to
avoid lawsuits as a close second. Avoiding government ﬁnes and penalties and the desire
to reduce their subjective probability of being caught out of compliance were tied for
farmers for third place. As highlighted in Chapter 6, this desire suggests that with

considerable uncertainty, farmers will expend considerable effort.

Table 6: Perceived Criteria Necessary to Reduce Uncertainty

1. Regulators must clearly tell farmers what is expected of them

2. Clear guidelines should be given to farmers on what they can do to comply with
the total maximum daily load process

3. These guidelines should be comprised of common sense practices, they should
be affordable, they should not be mandatory, allowing farmers to do what they
think is best for their operation

4. Farmers want to know the penalties for a violation of the TMDL process

5. Farmers want assurance that if they are in compliance with a TMDL, they will be
safeguarded against future regulatory changes and local zoning requirements

6. If they have to make any necessary investments both in terms of time and money,
farmers want assurance that the water quality regulations will result in improved
water quality outcomes

7. Farmers must not be punished for past environmental investments that they have
made in good faith

8. Ideally, the regulations would allow farmers to self-monitor their performance so
that they can clearly see the results of their management practices and make any
necessary changes in a timely manner

 

 

58 The Generally Accepted Agricultural and Management Practices (GAAMPs) speciﬁed under the
Michigan Right-to—Farm Act could reduce some uncertainty for farmers since they specify voluntary
approved manure management practices. Yet interestingly, few participating farmers mentioned the
GAAMPs as an existing set of voluntary guidelines that can help them ﬁx any known problems, and/or
reduce regulatory uncertainty by reducing the possibility of pollution runoff.

112

In light of the recent lawsuit notices given by the Sierra Club against ﬁve
Michigan dairies — all of whom had discharges as recorded by Michigan public agencies

— it is not surprising that farmers are anxious to avoid third-party lawsuits.

Table 7: Ranking of Reasons Farmers Would Comply with TMDL

 

 

Reason Importance
I need to protect our water resources for future generations 1
I want to avoid lawsuits 2
I want to avoid penalties and ﬁnes 3
I want to reduce the chances of beirg caught out of compliance 3

 

Most interviewed farmers were more wonied about lawsuits than government ﬁnes
because they believed a lawsuit could drag on and potentially put them out of business,
whereas government ﬁnes typically will not. Additionally, most participants felt that the
state government is more willing to work with farmers, unlike third-parties such as

neighbors and citizens’ groups.

Farmers were asked if they would be prefer to hire consultants to assist in
compliance. Feelings were mixed regarding how helpful consultants could be. The main
concern was whether consultants would “stick their necks out” and assume responsibility
for what they told farmers to do. Many interviewed farmers felt that whoever provides
technical information, whether it be a government agency person or a private consultant,
should take on some legal liability responsibility, since they have advised farmers on how
to come into compliance. Understandably, farmers want to be able to shift their liability
transaction costs onto others. This concern illustrates the relevance of lawsuits to
farmers’ preference for compliance measures, as mentioned in Chapter 6. A technical
sheet increases their certainty, causing a lessening of their subjective probability of being

sued, shifting their marginal penalty away from the Y-axis — corresponding to them

113

exiting the industry due to bankruptcy — down towards the X-axis. The ability to shift
liability costs to the regulator appears to be the foremost beneﬁt to farmers of the

technical sheet by reducing their abatement effort and associated costs.

Further evidence for the desire for certainty arose within focus group discussions.
For instance, many participants felt that if a farm is found to be in violation of the TMDL
standard, the farmer should be told what the problem is, told how he can ﬁx it, be
provided with adequate technical assistance to tailor a solution to his farm, and be given
an adequate amount of time to ﬁx the problem. Otherwise, it would be unfair to penalize
the farmer. In essence, farmers again indicated a desire for a technical sheet of approved
practices to be provided. This technical sheet is a proxy for compliance certainty, and was

preferred to a ﬂexible environmental performance standard.

Even farmers who simply did not trust any type of approved practice put forth by
government regulators — since they felt such efforts have been ineffective and expensive
in the past —— did want certainty with respect to compliance. One farmer said he wished
there were accurate measures of pollution runoff. Then, if there was a meter at the end of
his ﬁelds that could clearly tell him whether he exceeded his phosphorus limit and how

much he had to reduce his efﬂuent, he would have no need for an approved list.

Well ...if ...there ’s a meter there right next to the creek that ’s telling us
that we got to reduce 10%, so when the meter dips 10% we know where it
is. So I guess I trust my own ideas as much as I would anything the
government is going to come up with... Those approved lists are going to

strap you to more expense. Where if you did it on your own and the meter

114

is there at the other end that [tells you that] you ’ve dumped 10% then I’d

be confident [in my own judgment].

This farmer wants a technology — a meter — that provides him with compliance certainty
and the ability to limit his liability exposure, while enabling innovation. This statement
suggests that this farmer wants a ﬂexible efﬂuent standard, provided he has timely and
high quality “feedback” information regarding his efﬂuent in relationship to his efﬂuent

standard.

7.3.2 The Paradoxical Desire for Design Standards and Flexibility

Paradoxically, the interviewed farmers expressed a strong desire for certainty, but
also wanted ﬂexibility in the regulations. On the one hand, they wanted clear guidelines
for what they must do to be in compliance and be protected against lawsuits, but on the
other hand, they wanted these guidelines to be ﬂexible (Table 6, point 3). They did not
recognize the trade-off between ﬂexibility and compliance certainty illustrated in Figure

2, Chapter 3, reproduced on page 107.
This evidence from focus groups and the rankings in

Table 7 illustrate the farmers’ desire for certainty and a technical list, and it lends
evidence that farmers want to move from point A (i.e. a ﬂexible ambient performance
standard) to point B (i.e. a design standard-based environmental regulation) in Figure 2,
reproduced on page 107. The loss in ﬂexibility is traded-off against the ability to shift
legal liability onto the regulator. The liability rules speciﬁed by the TMDL was also a

concern for the focus group participants.

115

7.3.3 Equity Concerns

The farmers in the focus group discussions felt that the regulations, monitoring
efforts by the regulators, and enforcement of the water quality rules should not be

arbitrary. As one farmer noted:
“You want to see [the rules] be fair to everybody "

For example, participants stressed that monitoring of lakes and streams by state
regulators should be fair and should reﬂect reality. It was suggested that water samples
should not be taken during extreme storm events, and samples of every lake or stream
should be taken at several places, and at different times of the year. By using this method,
it is less likely that a farmer using effective manure management practices will be
wrongly accused of discharging. This ﬁnding provides evidence that farmers do not want
a ﬂexible ambient standard, where all farmers are penalized if the speciﬁed ambient
standard is exceeded, since it contradicts their sense of fairness. The participants felt that
those who undertake efforts to reduce efﬂuent from their farm should be given credit, and

those who do not should be penalized.

From the focus groups, it appears that farmers want the TMDL process to be
implemented with a negligence rule. With a negligence rule, the farmer is penalized only
if he fails to comply with the “due standard care.” Thus, a farmer who complies with the
TMDL process in good faith will not be ﬁned. In addition, these farmers’ desire for a
negligence rule and sense of fairness implies that they want regulators to measure the
efﬂuent from each farm. State regulators need to collect information from farmers about
their management practices in order to ensure that only farmers who do not undertake

sufﬁcient abatement efforts — as speciﬁed by the compliance measure — are penalized. In

116

essence, these farmers would like to see the TMDL process implemented as an efﬂuent
standard. But this is contrary to the current state of affairs in Michigan, where regulators

are not actively gathering information regarding farmers’ practices.59

Farmers felt that they should not have to bear any penalties in addition to their
own compliance costs. Their cost concerns and level of abatement effort are illustrated in
Table 8. In the exit survey, farmers were asked to rate reasons why they would not
comply with a total maximum daily load process. Cost considerations ranked amongst the
top three in terms of importance. If costs and fairness are important issues that affect
whether farmers will follow water quality regulations, then these discussions with
farmers imply that with high compliance uncertainty, some farmers may free ride rather
than expend a lot of effort in complying, especially if they feel that the regulator is not
really going to be monitoring and enforcing the TMDL standard. Thus, a comer solution

as discussed in Section 6.6 may be a real possibility.

Table 8: Weighted Ranking of Reasons Farmers Would not Comply with TMDL

 

 

Reason Weighted
Ranking Value
1. I have to carefully consider the costs of suitable water protecting 75
practices
2. Current milk prices affect the water-quality related practices that I 72
can undertake
3. I have to consider the costs of a consultant because of ﬁnancial 57
constraints

 

 

59 Regulators do visit farms if a nuisance complaint is ﬁled, but this data is gathered in response to a
complaint. Regulators do not otherwise know what choices farmers are making with respect to pollution
abatement.

117

CHAPTER 8

Summary and Conclusion

This analysis indicates that ﬂexible environmental performance standards may not
be cost-effective when applied to nonpoint sources of pollution when compliance
uncertainty results in signiﬁcant transaction costs, thereby contributing an improved
understanding in the economic literature of the limitations of ﬂexible environmental
performance standards. This chapter summarizes the contribution of this research to the
literature while answering the two research questions, outlines the policy implications,

discusses the caveats of this research, and suggests areas for future research.

8.1 Contribution to the Literature

The environmental economic literature has contended that ﬂexible performance
standards are potentially more cost-effective than design standards and thus are preferred
by economists. They have the potential to be more cost effective because, as noted by the
Porter Hypothesis, the ﬂexibility makes induced innovation and innovation offsets
possible. Innovation offsets may reduce compliance costs for regulated ﬁrms, and they
potentially reduce search and informational costs for the regulator since it no longer has

to determine and specify a design standard for regulated ﬁrms.

Yet, the economic literature that promotes ﬂexible performance standards does
not sufﬁciently account for the effect of compliance uncertainty and its resulting
transaction costs in assessing the cost-effectiveness of ﬂexible environmental
performance standards in controlling nonpoint sources of pollution. Compliance
uncertainty, due to imperfect information and the stochastic nature of nonpoint source

pollution, results in transaction costs that can overcome the beneﬁts of ﬂexible

118

environmental performance standards when they are applied to nonpoint sources of

pollution.

Regulators suffer from compliance uncertainty due to unobservable efﬂuent,
incomplete knowledge on their behalf with respect to understanding the fate and transport
of efﬂuent from nonpoint sources and its stochastic characteristics, and because farmers
can behave opportunistically due to asymmetric information between many actors and the
regulator, resulting in moral hazard. Farmers themselves suffer from uncertainty with
respect to whether their actions or effort will be deemed sufﬁcient for compliance, and

they are also uncertain as what other farmers are doing.

For an ambient compliance measure, the actions of other farmers may also affect
whether the farm is deemed compliant. These transaction costs resulting from compliance
uncertainty affect the beneﬁts that farmers receive through induced innovation, and thus
affect the likelihood that an innovation will occur. As highlighted in Chapter 6, in their
desire to minimize costs, both the regulator and farmer will prefer to reduce their total

expected costs resulting from compliance uncertainty.

This study examined the effect of uncertainty and its resulting transaction costs on
the cost-effectiveness of ﬂexible environmental performance standards to curb nonpoint
source pollution. This examination was guided by an ex ante analysis of the
implementation of total maximum daily loads by Michigan regulators to control
phosphorus and the compliance response by dairy farmers. The following two key
research questions were examined to give insight into the differential incidence of

farmer’s costs under different compliance measures:

119

 

 

1. How might compliance uncertainty affect the effort level of Michigan
livestock farmers, their costs, and their benefits from induced

innovation?

2. Can ﬂexible environmental performance standards ensure induced
innovation in the context of compliance uncertainty, and what are the

implications for the Porter Hypothesis?

The graphical models used in this paper show that differing levels of cost-savings
due to innovation result from various levels of compliance uncertainty with a ﬂexible
performance standard. In the model, the TMDL process is implemented alternatively as
an ambient standard, an efﬂuent standard, and a design standard to investigate the
abatement efforts and costs under each approach. The regulator ensures compliance
through monitoring and enforcement efforts and levying a penalty on those who do not
comply. As predicted, this analysis of the total maximum daily load process reveals that
transaction costs due to compliance uncertainty affect the effort level of farmers, their
costs, and their beneﬁts from induced innovation. Essentially, the greater the compliance
uncertainty with higher resulting transaction costs, coupled with a high expected marginal
penalty, the greater is the control effort by the farmer, and the greater are his costs to

comply with the TMDL process.

The analysis reveals three main lessons: 1) compliance uncertainty coupled with
high penalties provides farmers important additional incentives to innovate by allowing
them to adjust their level of effort, thereby reducing their expected penalty costs, 2) both

the regulator and farmer prefer the ﬂexible efﬂuent performance standard to the ﬂexible

120

ambient performance standard because compliance uncertainty is reduced60 — the
regulator has more information about farmers’ actions with ﬂexible efﬂuent standards,
reducing moral hazard — resulting in a lower expected marginal penalty, and lower total
expected costs for the regulated farmer, and 3) the uncertainty-reduction effect from an
innovation that decreases compliance uncertainty — shifting down or rotating the farmer’s
expected marginal penalty curve — results in lower effort and lower total expected costs
for the farmer, so farmers will choose those alternatives that will affect compliance

uncertainty since their costs are lower.

What are the implications for the Porter Hypothesis which predicts that ﬂexible
environmental performance standards will provide incentives to regulated ﬁrms to
innovate thereby reducing their compliance costs? In the case of farms, characterized as
nonpoint sources of pollution, compliance uncertainty can exacerbate the possibility of
lawsuits. That is, in such a situation, farmers may face liability issues that are not due to
their own management efforts. The focus groups showed that farmers want to avoid
lawsuits since they expose personal assets to loss. Consequently, farmers may prefer to
not invest in research and development efforts to reduce their compliance costs — that is,
they may not innovate — but would prefer to shift liability costs to the regulator by
demanding that the regulator implement design standards. If this were the case, then the
Porter Hypothesis would not hold since innovation is precluded when farmers follow a

design standards as discussed in 6.5.

 

60 Additionally, the ﬂexible efﬂuent performance standard may be perceived as being more fair and
equitable, and therefore more acceptable.

121

 

8.2 Policy implications

The total maximum daily load process poses a challenge to both regulators and
farmers because they have not been implemented before. Regulators in Michigan — and
throughout the country — have little experience in terms of technical expertise, model
building, data gathering, and monitoring and enforcement. Farmers too have faced few
environmental regulations historically, while smaller farms, which were the subject of
this study, have by and large never been subjected to consistent water quality regulations.
As such, they do not have a great degree of technical skills to deal with any agro-

environmental policy, let alone a TMDL process.

If regulators implement a TMDL process in the short-run with a technical sheet,
they will reduce their institutional demands; in fact, by examining practices as opposed to
examining actual environmental outcomes, little institutional change will be necessary.
The Michigan Department of Environmental Quality (MDEQ) has essentially been
conducting design-inspections for the past 30 years in its efforts to monitor point sources
that have National Pollutant Discharge Elimination System permits. Thus, an approved
list of practices requires few institutional changes by the MDEQ. Moreover, although
state regulators have limited budgets for developing a list of approved technical practices,
Michigan’s Generally Accepted Agricultural Management Practices (GAAMPs) can be

adapted as required, thereby reducing regulators’ program development costs.

The use of this list, however, precludes any induced innovation from occurring.
Nevertheless, the technical sheet can allow some degree of innovation if updates with

farmers’ and researchers practices are allowed, after these practices have been validated.

122

Regulators will have to continue working with researchers to better understand the link

between management practices and environmental outcomes.

Regulators can encourage dynamism — by incorporating ﬂexibility within a
TMDL process — through an adaptive approach. This approach is consistent with NRC’s
recommendation that TMDL plans be employed in an adaptive manner (NRC 2001).
Regulators may begin implementing the total maximum daily load process in the short
run by providing a technical sheet to farmers. But as they develop expertise in
implementing TMDL standards, regulators should work towards a ﬂexible standard to
allow for induced innovations and possible innovation offsets. The high degree of
compliance uncertainty and resulting transaction costs coupled with farmers’ equity
concerns make a ﬂexible ambient standard unattractive, leaving ﬂexible efﬂuent

standards as the remaining option.

The regulator will have to invest in reliable modeling, data gathering from farm
sites, and monitoring efforts to implement a ﬂexible efﬂuent standard. The regulator must
invest in research and extension activities to assist farmers in better understanding the
link between their practices and environmental outcomes. The trade-off is that the greater
ﬂexibility exposes farmers to lawsuits, where their personal assets are exposed to
potential loss. Understandably, this situation ensures farmers’ personal attention. But if
the regulator invests in modeling efforts and gathers high quality data from farm sites,
then this possibility can be reduced. With ﬂexibility and assistance, the regulator can
provide farmers with the incentive to undertake efforts that result in induced innovation
and, perhaps, in innovation offsets. Consequently, over time, with the right regulatory

measures, the Porter Hypothesis may hold in nonpoint source pollution control.

123

 

Essentially, farmers will choose those incentives that will allow them to reduce
their expected marginal penalty. With a ﬂexible standard, if farmers believe that their
innovation will reduce the probability of them being penalized, say by having a
signiﬁcant positive effect on the quality the impaired body of water, or if the regulator
responds to induced innovations by reducing the penalty, then they will prefer ﬂexible
performance standards, and will innovate, possibly resulting in innovation offsets. This
preference, however, depends upon the degree to which they believe the marginal penalty
will shift. For example, if the regulator invests resources in actively monitoring the
efﬂuent and responds to innovations by reducing penalties and informs the farmer of this
reduction, then this incentive will increase the probability of induced innovations and

innovation offsets, improving the relevance of the Porter Hypothesis.

Regulatory actions like enforcement depend upon political leadership, and this
changes over time. If state regulators do not provide the correct incentives to farmers,
then resources will not be allocated cost-effectively, neither by the regulator nor the
farmer. In fact, this model shows that with a high degree of uncertainty coupled with a
high marginal penalty, farmers over-invest relative to the case of perfect informaiton, and
may do so just to avoid lawsuits. Without consistent regulatory action, such as
monitoring and enforcing the total maximum daily load process, third-party citizen
groups will continue to be the defacto enforcer of the federal Clean Water Act, forcing
farmers to remain focused on efforts to improve water quality, but at a cost that is greater

than necessary.

124

 

8.3 Caveat for research

There are two main caveats to this research. The ﬁrst caveat is that this analysis is
an ex ante analysis. It does not analyze actual compliance measures that have been
implemented by Michigan nor responses by farmers. Nevertheless, it does provide some
information to both regulators and farmers as to what considerations are pertinent in their
decision-making process. At the time that this study was undertaken, there was little

information exchange between the state regulator and the farming community.

A second caveat is the data from the focus groups. These data were not
statistically representative of all dairy farmers in the state, so only qualitative results
could be presented. Yet, these results can provide a guide to a more quantitative analysis
once total maximum daily load process has been in effect in the state and data can be

collected from farmers.

8.4 Future research

The ﬁndings from this research suggest a number of areas for future research,
such as the role of liability, the need for a dynamic analysis, and ex past studies. The
overwhelming concern with respect to legal liability is an area to be explored both as a
motivator -— indeed as a substitute for regulatory actions — and as a risk to be minimized.
With little real information, even from extension agents, regarding the requirements
needed to avoid lawsuits for violations of the Clean Water Act, farmers fear being put out
of business by lawsuits. The model in this paper ﬁnds that high penalties and their
concern of being put out of business by lawsuits may cause farmers to over-invest in

abatement efforts.

125

The economic literature needs to be more aware of the role played by imperfect
information that leads to compliance uncertainty when advocating ﬂexible ambient
performance standards. Flexible performance standards may be more appropriate for
some point sources of pollution, where the pollutant level is observable and easily
quantiﬁed. This ease of measurement eliminates most uncertainty as to whether point
sources are in compliance with the environmental regulation. But with nonpoint sources
of pollution, compliance uncertainty and its resulting transaction costs indicate a need for
additional economic research into uncertainty relating to the implementation of ﬂexible

environmental performance standards to control nonpoint sources.

Clearly if the total maximum daily load process is to achieve its objective,
considerable research needs to be undertaken to investigate ways in which the transaction
costs resulting from compliance uncertainty can be reduced, for both the regulator and
farmers in Michigan. For instance, regulators and dairy farmers need to be able to link
practices with resultant water quality outcomes in a cost-effective and equitable manner.
Currently there is insufﬁcient information and knowledge to target enforcement on the
major sources of pollution, including farms, and to tailor solutions to their needs.
Monitoring of water bodies and enforcement by regulators must be undertaken; research
may help to reduce these ex post transaction costs for regulators. Additionally, research to
improve fate and transport models of nonpoint source is needed by both regulators and
farmers. Research efforts can help ﬁll these knowledge gaps to ensure that the beneﬁts of
ﬂexible environmental performance standards, such as total maximum daily loads, are

realized, thereby ensuring continual water quality improvements in Michigan.

126

 

Data from these research efforts can lead to a richer analysis that incorporates the
strategic actions of regulators and regulated ﬁrms when information is imperfect due to
asymmetries and poor scientiﬁc, technical knowledge, and imperfect monitoring and
enforcement. A dynamic model that accounts for all these characteristics may be able to
more fully answer the question: Can ﬂexible environmental performance standards

improve environmental outcomes?

127

APPENDIX A

Survey with Regulators

128

Issues for Regulators

 

1) Implementation of TMDL

2)

3)

a) When will they be implemented here in MI?

b) How is the threshold being speciﬁed?

0) What will compliance mean?

(1) How will compliance be determined?
- E.g. +/- 10% 80% of the time?
- How ﬂexible will they be?
- How will they know who is in compliance?
- How will they be able to detect if the threshold is not being met?
- How will monitoring be carried out? Agency personnel, or say if there is a

complaint?
- Will farmers be given a suite of technologies, and what practices/behavior is
considered to be “in compliance”

- Will discharges be measured from ﬁelds?
- Will farmers be given permits?

e) How strong will enforcement be?
- What’s the probability of being caught? Of being penalized?

f) If the TMDL is not being met, what is the consequence?
- Mandated technologies? If so what?
- Do they know what the costs will be?
— How do they follow-up? Monitor?

g) What are EPA requirements?

h) Does everyone have to be in compliance?
— Dividing TMDL evenly by all the people? or?

i) Are they following a model from elsewhere (e.g. Oregon?)

j) Do they have an idea of how farmers will respond?

k) What do farmers think of TMDLs?
- Have any been involved in the public comment phase?

1) What incentives are they providing to farmers to meet the TMDL?

m) Cost-sharing for adoption of abatement technologies to meet TMDL?

Third Generation BMPs

a) What are they? Where will they come from?

b) What are the costs of implementing these BMPs?

c) How much P does each BMP reduce? %, quantity ?

Legal issues

d)
e)
0

Who can sue whom?

What is the connection between CWA and Michigan Right to Farm Guidelines?

Does liability insurance exist? Are farmers using it?

129

 

APPENDIX B

Materials for Farmer Focus Groups

130

 

Environmental Performance Standard Script

Presented by Lorie Srivastava
Graduate Student, Dept. of Agricultural Economics

Introduction
Welcome. I would like to thank all of you for participating today. I know this is

an extremely busy time of the year, so I really do appreciate you taking the time to come.
Let me ﬁrst introduce myself and explain the purpose of this discussion. My name is
Lorie Srivastava, and I will serve as the facilitator for this meeting. I am a Ph.D. student
in the Department of Agricultural Economics at Michigan State University. I would also
like to introduce Corey Risch. Corey is a Masters student in the Dept. of Agricultural
Econorrrics. She is from a dairy farm in Ingham county. They milk 175 cows. Corey will
be making notes of what is said during this meeting, so that we can see keep track of our
discussion. In a moment I will describe my role in more detail, but ﬁrst I would just like

to say a few words about why we invited you here.

Purpose

Today’s meeting is part of my Ph.D. research. As part of my research, I am
looking at how Michigan dairy farmers might be affected by federal water quality
regulations that are being implemented in Michigan under the Clean Water Act, and how
important issues like third-party lawsuits, and pollution insurance affect farmers. I will be
holding 6 discussions like this in the 6 largest milk producing counties in Michigan to
gather the concerns and opinions of dairy farmers. I will be speaking with farmers from a
variety of farm sizes — small, medium, and large, people who intend to stay in dairy for

the next 15-20 years, and people who will be retiring soon. I want to talk to a variety of

131

dairy producers because the industry is made up of different kinds of operators, who are

making different kinds of decisions.

The intent of today’s roundtable discussion is to ensure that your knowledge and
concerns are incorporated into my research. Does this mean that what is said here will be
written up and forgotten in the library at MSU? No! After all the roundtable discussions
are held, I will summarize what you tell me and disseminate the information to various
policy makers so that they can be informed of dairy farmers’ perspectives when they

implement water quality regulations in Michigan.

I should mention that I do not work for MDA, DEQ or the EPA, and have no
plans to do so. As I said before, this work is for my Ph.D. research. I want to admit right
up front that I am not an expert in dairy, which is why I am asking you for your help with
my research. My research is funded by the Great Lakes Protection Fund, which provides
funds for a variety of research issues pertaining to the Great Lakes. I would also like to
mention that this has nothing to do with the Michigan Agricultural Environmental

Assurance Program (or MAEAP), in case any of you were involved in those focus
groups.

Lorie: keep in mind the policy makers are MDA, DEQ — and they can in turn inform

EPA

I want let you know that I am not doing this research because I think water
regulations for agriculture are a good or a bad thing, or that mandatory water quality
regulations are needed for agriculture. Rather, I believe that there has been a lot of
research on voluntary environmental programs, but less on potentially mandatory

environmental programs like those that we will discuss today. And, as far as I know,

132

 

there has been almost no research done that directly asks dairy farmers what they think
about possible mandatory water quality programs, what they think the costs will be, how
they think their operation will be affected, how their production decisions might change,
and what they think the beneﬁts are. The water quality regulations that are discussed in
the LSJ article have not been implemented in Michigan yet because it is still being
designed and debated, so this is an opportunity for farmers to let their concerns be

known, which I think is important.

I am here to facilitate this discussion and to listen to you — I am not here to judge
you in any way. I just want to hear directly from you what concerns you about upcoming
water quality legislation, lawsuits, and how you believe these will affect you. Let me give

you some background.

Background

It looks like farmers — particularly livestock farmers — are faced with increasing
regulatory pressures because of agricultural contributions to water pollution, and that
such pressures will become stronger with time. The regulatory pressure is in response to
the declining acceptance by the general public of many current farming practices because
of perceived water pollution from agriculture. The cover story in the Aug. 20th issue of
the Lansing State Journal talks about agriculture’s contribution to water pollution, and
how new legislation will change it. The implication is that runoff from agriculture has to

stop because runoff, along with other sources, is causing Michigan waters to be unsafe.
Lorie: hold up newspaper

An example of this public concern recently happened here in Michigan. As you

may know, on July 25th of this year, the Sierra Club gave notice that it is suing four

133

 

Michigan dairies under the Clean Water Act.61 The Sierra Club contends that these
dairies have repeatedly discharged manure into Michigan’s waterways, which is a

violation of the federal Clean Water Act”.

In Michigan, it is commonly thought that using the Generally Accepted
Agricultural Management Practices (GAAMPs) speciﬁed under the Right to Farm Act
(RTFA) protects farmers from environmental complaints and from too many inspections
from state regulators. It is, however, possible that a farmer using GAAMPs could have a
discharge of wastes into surface waters, and if so, they could be penalized. So, RTFA

does not actually protect farmers from violations of the Clean Water Act.

So where does that leave us? What will future water quality regulation for
agriculture look like? This is a big unknown right now. It is likely that in the future, at
least those farmers with signiﬁcant potential for pollution discharges will increasingly be
asked to change management practices so as to further protect water quality for drinking
water, ﬁsh habitat, and to reduce sediment runoff. Water quality regulations may
continue as in the past with voluntary programs with incentives to get farmers to
participate, or instead it might be composed of mandatory requirements with required
management practices. A third possibility is that agriculture may be governed by what

policy makers call environmental performance standards.
Lorie: hold up visual — a ﬂow diagram

So, what are environmental performance standards? It is a standard that sets a

limit on how much of a certain pollutant can enter a surface body of water, by surface

 

61 In case it is asked, the 4 dairies are: are Bruinsma Farms, River Ridge Farms, Walnutdale Farms, and
Bradford Dairy Farms - Kent county, Ottawa County, others??

62 Lorie: Sierra Club is just making the point to explain the need for federal enforcement

134

body I mean a lake, or a stream, or creek, something on the surface, as opposed to
groundwater. Let me illustrate with a hypothetical case of phosphorus in a lake. An
environmental performance standard would set a pre-determined amount of phosphorus
that can legally exist in the lake. If a state regulator (such as MDA/DEQ) were to check
the phosphorus level, and they found that the amount of phosphorus in the lake was
below the allowed standard, then no further action is taken. If, however, the regulator
tests the water and ﬁnds that there is more than the allowed level, then there would have
to be a cutback in the amount of phosphorus entering the lake. I am using phosphorus as
an example because it promotes rapid algae growth, and when the algae dies and decays,
the oxygen in the water is used up, causing ﬁsh kills and other creatures to die. It is the

main concern from agricultural runoff, along with sediments. Any questions?

One type of environmental performance standard is a total maximum daily load
process, or so called TMDLs (some people pronounce it timdls, others TMDLs, I will say

TMDLs). How many of you have heard of total maximum daily loads, or TMDLs?

In order to direct the discussion, I would like to focus on total maximum daily
loads since it is the only example of an environmental performance standard in
agriculture. Also rules are being developed for it, so I am interested in how farmers may
react to it, if it is implemented. First, let me explain a total maximum daily load. A total
maximum daily load is a process that is Speciﬁed under the federal Clean Water Act. The
current total maximum daily load process calls for each state to set an environmental
performance standard for every surface water body within its jurisdiction (this has
already been completed in Michigan). Different water quality standards are set for waters

that are used for different purposes, such as drinking water source, recreation, ﬁshery,

135

 

etc. The EPA is letting states decide whether to use mandatory or voluntary approaches to
implement any clean up plan. They have 10 years to do so. I should add that government
authorities expect all sectors of the economy to be regulated with TMDLs, not just

agriculture.
Lorie: Go through the visual

The total maximum daily load process calls for the state to regularly check each
water body and see if it meets its environmental performance standard. As an example,
can you look at the handout of the lake that I have passed around? There are two lakes,
which are surrounded only by farms —- no waste water facilities, or urban golf courses, nor
anything else, are near the lakes. The smaller lake meets its designated water quality
standard, and so does not exceed its allowable level of phosphorus. Hence Farm B does
not have to change any management practices. The larger lake, however, does not meet
its water quality standard. There is too much phosphorus in this lake —— so as you can see
in this example, Farm A has to cut back its phosphorus by 15%, farm B and C by 40%,
and Farm D does not have to change any practices. These amounts would be determined
by a regulator, along with a timeline for when the changes must be completed.

Questions?

Since total maximum daily loads have not been implemented in Michigan yet,
how it will be speciﬁed is not known. There are a few options: 1) they may be
implemented so that only major polluters have to reduce their pollution, not everyone. In
this case, only those farms that are a major source of water pollution will have to reduce
their runoff, not all farms; 2) farmers may be given a lot of ﬂexibility to meet a needed

reduction. So it may be up to the farmer to ﬁnd a cost-effective way to bring the farm into

136

compliance — for example, the farmer can change hauling practices, build additional
storage, etc. 3) Or total maximum daily loads can be implemented so that farmers have to

follow speciﬁed practices — they can not choose their practices.

In July of this year, the EPA issued a ﬁnal rule on how to implement the TMDL
process. Since August 1999, the EPA has considered over 34,000 comments on changes
that it proposed to the TMDL process. Here in Michigan, the DEQ and the EPA are going
back and forth on how TMDLs will be implemented, with no resolution. A major bone of
contention is the issue of discharge permits. The EPA is pushing DEQ to issue discharge
permits to agricultural operations, as called for in the Clean Water Act63. But the DEQ
does not grant discharge permits to agriculture, since technically agricultural enterprises
are not allowed to discharge into Michigan waterways at all. So the DEQ is arguing with

the EPA about discharge permits.

What do TMDLs have to do with lawsuits like the ones being ﬁled by the Sierra
Club? Third-party lawsuits against ﬁrms (such as farms and processors) are used to put
pressure on regulating agencies like the EPA and DEQ to implement sections of the

Clean Water Act that have been ignored until now, like total maximum daily loads.

I would like to now begin the questions, but want to point out that there are no
right or wrong answers to the questions that we will be discussing today. I just want to
know what you believe are the important issues for your farm with regard to water

pollution.

If you are wondering why you are here — you have been selected from a

nomination list submitted by extension dairy agents. I’m not trying to pick on you or

 

63 this is for conﬁned operations of over 700 mature dairy cows

137

anything! Finally, I know I can not fully compensate you for your time, I do want to

recognize your contribution, so you will be paid $100 for your time today.

Procedure

Before we begin our discussion, I would like to explain a few things.
There are several things that I need you to do to help improve my research:

1. I would like to hear from everyone, since I expect that there are a variety of
Opinions in this room, some in disagreement, some in agreement. All your
opinions are valid —they’re just based on different experiences — so I am
genuinely interested in hearing what all of you have to say. Please don’t worry
about offending me. You won’t! Although a roundtable discussion format allows
us to have an open discussion, and a chance to bounce ideas off each other, we
have a limited amount of time to get through a full list of questions. So there may
be times when I ask you to summarize or shorten your remarks to make sure that
everyone has a chance to give their input. If I change the subject, it’s not because
I’m not interested in what you have to say. It just means that we have to move

forward in order to be done by noon.

2. I hope that you will feel comfortable enough to share your honest opinions. Your

identities will remain conﬁdential, through the following 3 means:

0 no names will be used in the analysis

0 though I will be taping our conversation so that I can have an accurate
record of what is expressed, I will remove all references to peoples’ names

and replace them with number codes when I transcribe the tape

0 I am asking that you treat comments made today by other participants as

conﬁdential as well

138

o I would ask that you be honest in your answers. I will take every
precaution to ensure the conﬁdentiality of what is said here today. No one
will be able to trace back any comments that you make to you. I will not
be recording your names in my dissertation. I only need your names so

that the accounting folks at MSU can issue you a check.
. I hope that you will feel comfortable enough to share your honest opinions. Your

identities will remain conﬁdential, through the following 3 means:

. I would ask that everyone try to speak up and speak clearly so that the tape can

pick up your comments.

. At the end of our discussion, I would ask that you complete and leave a short

questionnaire that should only take about 10 minutes.

. Does anyone have any questions before we begin?

139

Focus Group Questions64

Attitudes towards water quality issues

1. What kinds of production decisions are inﬂuenced by water quality

considerations, e.g. day to day decisions, planning for the season, or when

planning an expansion?
a) Probe one individual or groups of individuals: If so, in what way?

b) Probe one individual or groups of individuals if none: If not, why not?

What comes to mind ﬁrst?

2. In a good ﬁnancial year, how high a priority would improving your manure

handling system be? Top 3, top 5?
a) What affects how high a priority it is for you?

3. In general, do you feel that serious water quality problems stem from dairy farms

in Michigan? Why or why not?
Concerns about water quality regulations
Lorie: Hold up TMDL visual

4. Thinking of your farm, if you were to make an investment that reduced runoff
(such as equipment/facilities and additional training for you and your employees),
would there be any advantages for making these investments and being in

complete “compliance” or even “over-compliance” with a TMDL? If so, what

would be the main advantage?

 

64 Due to time constraints, not every focus group answered every question.

140

5. Now I am going to present you with a hypothetical scenario. Suppose that a state
regulator determines that a water body in your watershed is being polluted above
an acceptable standard that is set by the state. Suppose there is evidence that there
is animal manure and other dairy farm wastes runoff entering this water body.
Assume your farm is contributing to this pollution, like farm A. Would you be
more worried about being sued by a third party (like a neighbor or an citizen

group), or being ﬁned by the government, or would neither be a concern?
a) If so, which is the bigger concern? Why?

6. Actually, before we move on, have any of you heard of ﬁnes by DEQ and/or the
EPA being assessed against dairy operations here in Michigan? Were they large
ﬁnes? I know I can look it up, but does anyone know?

Lorie: hold up TMDL visual again

7. What do you think of TMDLs?

a) What is your biggest concern regarding a water quality performance

standard for agriculture, such as a TMDL process?
b) Do you think TMDLs are appropriate for reducing runoff?

Lorie: Show TMDL 2 visual:
c) If there is a TMDL process, should an agency assign responsibility to

exceeding a water body’s TMDL limit to an individual farm?

8. A discharge permit speciﬁes the steps that farms have to take to minimize water
pollution. As long as you follow the permit requirements, you are protected to

some extent against law suits. Do you think it is appropriate for the state of

141

Michigan to issue discharge permits to dairy operations. If so, which farms and
why?

9. If the TMDL process is implemented next month, and you are told by the state
regulating agency that you have to reduce phosphorus loadings by 10%. Say you
decide to follow the TMDL process as closely as possible, which of the following

would you do?

3) Would you evaluate your operation and devise new practices on your own

to comply with the TMDL and cutback phosphorus by 10%?

b) Say DEQ/MDA/NRCS develops a list of approved practices where you
could use those that are best suited for your farm and cutback phosphorus
by 10%. If you use the practices on this list, you would be protected from
lawsuits to some degree. Would you like to be given this list of approved

practices and apply those that are suitable to your farm?

c) Now suppose, that you come up with a way that can signiﬁcantly reduce
the runoff from your farm, so that you meet the allowed phosphorus limit
in a TMDL, but MDA and NRCS are unfamiliar with it, and it is not on
their list of approved practices. And let’s say that your practice is cheaper
to use than those on the approved list — would you use yours or the more

expensive approved practices? Why?

Pat said that I may need to clarify that a performance standard speciﬁes an outcome (1'. e.
control of P), and not which practices are used — since farmers are used to NRCS and

FSA checking practices

142

 

d) Would you rather hire a consultant? Why?

e) Probe: If you would hire a consultant, how much would you pay to ensure

that you are in compliance with the law?

10. I explained earlier that things are in ﬂux with regard to water quality laws both
federally and here in Michigan at the state level. Laws change, different people
are elected into ofﬁce, who knows where exactly it will all end up. Imagine you
are Farm A, do you think you would: 1) do something immediately, 2) wait for

things to settle down to do something, or 3) ignore it.
a) Probe those who say will wait: What are you waiting for and why?
How will regulations affect your farm

11. Now let's assume that an environmental performance standard such as a TMDL
process which I talked about earlier were implemented next month, and assume
your farm is assigned the responsibility of reducing, say phosphorus, any way you

can.

12. If the TMDL process is implemented next month, and you are told by the state

regulating agency that you have to reduce phosphorus loadings by 10%.

a) What would you do to comply with the TMDL process? In terms of:

management changes, new investments, seek out information?
b) Where would you seek assistance?

0) Would your efforts to comply with the TMDL process be different if it is

enforced by an honor system like now where you are not checked unless

143

there is a complaint, where there would be little monitoring by state
regulators, vs. a system that is enforced with periodic inspections? Which

do you prefer?

(1) Would it matter who did the enforcing, DEQ or MDA? Why and who do

you prefer?
Concerns about lawsuit uncertainty

13. Let’s return to the situation of the Sierra Club giving notice to sue those four

farms in Michigan. Do you worry about such lawsuits being ﬁled against you?

Why or Why not?

a) If so, what actions, if any, have you, or might you undertake to reduce the

likelihood of a lawsuit against you?

14. Do you think regulations like total maximum daily loads will have any effect on

lawsuits being ﬁled? How?

15. Do you think there will be more lawsuits like this one in the future — do you see

this as a trend?

a) Probe those who say yes: If you think there is a trend towards more

lawsuits, do you think you will change how you farm? How?

16. How many of you have special provisions in your insurance against a pollution

event such as a lagoon breaking that leads to a ﬁsh kill?
a) Why did you get it?

b) Do you think it is adequate for environmental pollution events?

144

 

0) Do you believe your special insurance will provide you with some

ﬁnancial protection against environmental lawsuits?

17. Do you have any further thoughts on the Sierra Club lawsuit that are relevant to

today’s discussion?

18. Finally, is there anything else you would like to tell me today?

Wrap-up
This completes our discussion today. Thank you very much for participating.

Before you leave, there is a brief questionnaire that I need you to complete. PLEASE DO
NOT PUT YOUR NAMES on it. We want your answers — like your comments here
today — to remain anonymous. I would like to thank you all for sharing your thoughts on

these questions.

145

 

Visual 1: What is an Environmental Performance Standard?

 

Environmental Performance Standard

Set standard/limit

Meets Standard? No further
acﬁon

 

Reduction necessary

 

 

 

146

Visual 2: Who is Responsible?

 

Total Maximum Daily Loads
Responsibility

         
 

Limit is exceeded

. .fﬁ‘""‘<f\(l..

aldlﬂlu

 

 

 

147

Visual 3: Allocating TMDL Amongst Farms

 

Total Maximum Daily Loads

   
   
 
 
   

 

Farm C

Farm A Reduce 30%

Reduce 10% Limit is exceeded

   

Farm B
Reduce 30% Farm D

Reduce 0%

    

Limit is not
exceeded

Farm E
No need to reduce

 

 

 

148

 

Water Quality $3
Producer Opinion E i f"
. , ,/
Exnt Survey i)”; l ’11,,”

., u A

Your concerns

1. How serious do you believe surface water (e.g. lake, river, creek) quality problems

are:
not at all somewhat serious very serious
serious serious

in the nation as a whole 1 2 3 4

in Michigan 1 2 3 4

in your county 1 2 3 4

from your farm 1 2 3 4

2. In your opinion, which of the following entities are the top three (3) contributors to
surface water quality problems in your county? Please choose and rank the top three

(3) only, with number 1 being the most responsible.

Responsible Entity Rank
Urban residents (e.g. lawn maintenance, golf courses)

 

Sewer overﬂow (from urban areas, into rivers during
storms)

 

State of Michigan (e.g. highway maintenance)

 

Waste water treatment facilities

 

Industry/manufacturing facilities

 

Row crop farmers

 

Swine farmers

 

Poultry farmers

 

Small dairy farmers (100 milking cows or less)

 

Medium dairy farmers (101-300 milking cows or less)

 

Large dairy farmers (301 milking cows or more)

 

Beef cattle farmers

 

Other livestock enterprises (e.g. sheep, horses)
Other:

 

 

 

149

3. How concerned are you about the following issues on your dairy farm?

Issue not at all somewhat concerned very
concerned concerned concerned

Creating/maintaining good 1 2 3 4
wildlife habitat on my property,
or property adjoining my
property
Controlling water and manure 1 2 3 4
run-off
Storage and disposal of animal 1 2 3 4
waste
Soil erosion control 1 2 3 4
Protection of groundwater 1 2 3 4
Protection of surface waters 1 2 3 4

150

g)

h)

i)

If you had to strictly comply with a mandatory TMDL process, to what extent do

you agree with the following motivations as reasons for you to comply with the

TMDL?

Motivation

I will do whatever the law requires

I care about my neighbors’ opinions of
my management style

I want to avoid penalties and ﬁnes
I want to avoid lawsuits

I need to have access to pollution
insurance

I need lower insurance premiums
because I can’t afford higher premiums

I need to protect our water resources for
future generations

I want to reduce the chances of being
caught out of compliance

I want to follow a list of approved
practices (approved by state and federal
regulators), even if the list is not
mandatory

I need to have access to consultants
because I need access to their
knowledge and technological expertise

strongly
agree

1
1

agree

disagree

strongly
disagree

4
4

From the list of reasons above, pick the one that represents the most important reason

you would comply with a mandatory TMDL process: Letter:

151

 

 

 

6. Suppose a mandatory TMDL process is implemented and you are expected to
comply with it, to what extent would you agree with the following motivations as

reasons why you would not comply:

Motivation Strongly Agree Disagree Strongly
agree Disagree
a) I have to carefully consider the costs of 1 2 3 4
suitable water protecting practices
b) I have many other things to take care of 1 2 3 4
that take priority over water quality
c) Current milk prices affect the water— 1 2 3 4
quality related practices that I can
undertake
d) I have to consider the costs of a 1 2 3 4
consultant because of ﬁnancial
constraints
e) I want to wait until everything is l 2 3 4

ﬁnalized before making any necessary
management changes or investments

7. Do you think the state should issue discharge permits to dairy farms in Michigan?

0 Yes 0 No 0 Don’t know
Information about your operation

8. What is the size of your dairy operation?

No. of milking cows: ' No. of heifers (birth to 1" calving):
No. of acres of tillable land for manure

application:

9. In what counties do you farm?

 

10. In what county do you live?

 

11. What is the ownership arrangement for your farm?

0 Sole proprietorship O Corporation - family
0 Family partnership 0 Corporation — non-family
O Partnership with non-family member 0 Other

 

152

12. How many non-farming households (unrelated to those living on the farm) are located

within 0.5 miles of your milking parlors?

Milking parlor 1: 0 to 4 households 5 to 10 households More than 10

O O households
0
Milking parlor 2: 0 to 4 households 5 to 10 households More than 10
O O households
0
Milking parlor 3: 0 to 4 households 5 to 10 households More than 10
O O households
0

13. In the last three years, how many times have you received a formal nuisance

complaint in which you have been contacted by MDA?

Never 1 to 3 times More than 3 times

0 O O
14. What manure storage method(s) do you use? (Choose all that are applicable)

 

Daily scrape Concrete Outdoor earthen Other (please specify)
and haul storage storage
0 O O
15. If you have storage, how many months do you have? month(s)

 

16. What type of application method(s) do you use? (Choose all that are applicable)

 

Surface Injection Irrigation Other (please specify)
application 0 O
O

17. How often do you test the soil on which manure is applied? (please mark only one

circle)

Never Once each 5 years Once each 3 years Every other year Annually
O O O O O

18. In which watershed is your primary dairy operation located (i.e. what is the name of

the closest river or creek to your property)?

19. Do not know: 0 or Name:

 

153

20. Do you have a creek or river that runs through your farm property or borders it?
0 Yes 0 No

If you answered NO, how far is the nearest lake, river, or creek from your farm?
miles

21. What type of slope does the majority of your land have?

 

0-1% (ﬂat) 2-5% more than 5%
O O O

22. Do your cattle utilize:

rotational grazing pastures conﬁned feeding facility both
0 O

23. Have you made any investments in the last 10 years in manure handling practices?

0 Yes 0 No
If you answered YES, how much has it cost you (approximately)? $

24. Have you made any investments in the last 10 years in improving neighbor relations?

 

0 Yes 0 No

If you answered YES, how much has it cost you (approximately)? $
What kinds of things have you done?

 

 

 

25. Which of the following water quality related management practices do you follow?

(check all that apply)

0 ﬁlter strips
0 animals are conﬁned all the time

Do you have a stream/creek/river adjoining your property? 0 Yes 0 No

If yes, do you fence off the stream/creek/river from your cattle? 0 Yes 0 No
C) control barn runoff (How? )
O Other(s)

 

 

154

26. Will you be retiring within the next 5 years? 0 Yes O No

If you answered YES, are you considering any of the following?
0 pass farm onto family/relative who will keep it in agriculture

0 pass farm onto family/relative who will keep it in dairy
O sell land to someone for non-agricultural purposes

155

REFERENCES

Adler, R.W. “Integrated Approaches to Water Pollution: Lessons from the Clean Air
Act.” Harvard Environmental Law Review 23 (1999): 203-295

Batie, 8.3. “Environmental Issues, Policy and the Food Industry.” L.T. Wallace and W.R.
Schroder, eds. Government and the Food Industry: Economic and Political Effects of
Conﬂict and Co-operation. Dordecht, The Netherlands: Kluwer Academic Publishers.
1997.

Batie, SS. and DE. Ervin. “Flexible Incentives for Environmental Management in
Agriculture: A Typology.” in F. Casey, A. Schmitz, S. Swinton, and D. Zilberman (eds.)
Flexible Incentives for the Adoption of Environmental Technologies in Agriculture.
Norwell, MA: Kluwer Academic Publishers. 1999.

Battle, J .B. and M.I. Lipeles. Water Pollution. 3rd edition. Cincinnati: Anderson
Publishing Co. 1998.

Boggess, W.G., G. Johns, and C. Meline. “Economic Impacts of Water Quality Program
in the Lake Okeechobee Watershed of Florida.” in SS. Batie, D.E. Ervin, and MA.
Schulz (eds) Business-Led Initiatives in Environmental Management: The Next
Generation of Policy? Proceedings of a Pre-conference Workshop to the AAEA
Meetings, Toronto, Canada. Department of Agricultural Economics. Special Report
(SR)92. Michigan State University. 1997.

Bromley, D.W. “Markets and Externalities.” in D.W. Bromley Natural Resource

Economics: Policy Problems and Contemporary Analysis. Boston: Kluwer Academnic
Publishers. 1986.

Cabe, R. and J .A. Herriges. “The Regulation of Non-Point Source Pollution Under
Imperfect and Asymmetric Information.” Journal of Environmental Economics and
Management 22(1992): 134-146.

Coase, R.H. “The Problem of Social Choice.” Journal of Law and Economics 3(October
1960): 1-44.

Driesen, D.M. “Is Emissions Trading an Economic Incentive Program?: Replacing the
Command and Control/Economic Incentive Dichotomy.” Washington and Lee Law
Review 55(1998): 289-350.

Environmental Law Institute. Enforceable State Mechanisms for the Control of Nonpoint
Source Water Pollution. Retrieved on May 1, 2002 from the World Wide Web:
http://www.epa.gov/OWOW/NPS/elistudy/execsum.html. 1997.

Gardiner, D. and RR. Portney. “Does Environmental Growth Conﬂict with Economic
Growth?” Resources 115(Spring 1994): 19-23.

156

Grifﬁn, RC. and D.W. Bromley. “Agricultural Runoff as a Nonpoint Externality: A
Theoretical Development.” American Journal of Agricultural Economics 64(August
1982): 547-552.

Hanley, N., J .F . Shogren, and B. White. Environmental Economics in Theory and
Practice. New York: Oxford University Press, 1997.

Heyes, A.G. “Making Things Stick: Enforcement and Compliance.” Oxford Review of
Economic Policy 14(Winter 1998): 50-63.

Hicks, J .R. The Theory of Wages. London: MacMillan, 1932, in J. P. Chavas, M. Aliber,
and TL. Cox. “An Analysis of the Source and Nature of Technical Change: The Case of
U.S. Agriculture.” Review of Economics and Statistics 79(November 1997): 482-92.

Horan, R.D., J .S. Shortle, and D.G. Abler. “Ambient Taxes When Polluters Have
Multiple Choices.” Journal of Environmental Economics and Management 36(1998):
186-199.

Horan, R.D., J .S. Shortle, and D.G. Abler. “Ambient Taxes Under m-Dimensional Choice
Sets, Heterogeneous Expectations, and Risk-Aversion.” Environmental and Resource
Economics 21(February 2002): 189-202.

Houck, O.A. “TMDLs, Are We There Yet?: The Long Road Toward Water Quality-
Based Regulation Under the Clean Water Act.” Environmental Law Reporter 27(August
1997): 10391-10401.

Houck, O.A. The Clean Water Act TMDL Program: Law, Policy, and Implementation.
Washington, DC: Environmental Law Institute, 1999.

Innes, R. “Regulating Livestock Waste: An Economic Perspective.” Choices, Second
Quarter, 1999: 14-19.

Jackson, P.W. “Lawsuits Targeting Farmers Called ‘Tip of the Iceberg’.” Michigan Farm
News. August 15, 2001. pp. 1,5.

Jaffe, AB. and K. Palmer. “Environmental Regulation and Innovation: A Panel Data
Study.” Review of Economics and Statistics 79(November 1997): 610-619.

Lemunyon, J. and TC. Daniel. “Phosphorus Management for Water Quality Protection:
A National Effort” in J. T. Sims (ed.) Soil Testing for Phosphorus: Environmental Uses
and Implications. Southern Cooperative Series Bulletin No. 389, SERA-IEG 17. North
Carolina State University. Raleigh, NC. 1998.

Mansﬁeld, E. Microeconomics: Theory and Applications. 6th Edition. New York: W. W.
Norton and Co. 1988.

McCann, L. and K.W. Easter. “Transaction Costs of Policies to Reduce Agricultural
Phosphorus Pollution in the Minnesota River.” Land Economics 75(August 1999): 404-
414.

157

McCann, L. and K.W. Easter. “Estimates of Public Sector Transaction Costs in NRCS
Programs.” Journal of Agricultural and Applied Economics 32(December 2000): 555-
563.

McCluskey, J .J . “A Game Theoretic Approach to Organic Foods: An Analysis of
Asymmetric Information and Policy.” Agricultural and Resource Economics Review
29(April 2000): 1-9.

Milgrom, P. and J. Roberts. Economics, Organizations and Management. Englewood
Cliffs, New Jersey: Prentice Hall. 1992.

Milliman, SR. and R. Prince. “Firm Incentives to Promote Technological Change in
Pollution Control.” Journal of Environmental Economics and Management 17(1989):
247-265.

National Research Council. Assessing the T MDL Approach to Water Quality
Management. Washington, DC: National Academy Press. 2001.

Norris, PE. and AP. Thurow “Environmental Policy and Technology Adoption in
Animal Agriculture” in F. Casey, A. Schmitz, S. Swinton, and D. Zilberman (eds.)
Flexible Incentives for the Adoption of Environmental Technologies in Agriculture.
Norwell, MA: Kluwer Academic Publishers. 1999.

Ondersteijn, C.J.M., A.C.G. Beldman, C.H.G. Daatselaar, G.W.J. Giesen, R.B.M. Huime
“The Dutch Mineral Accounting System and the European Nitrate Directive: Implications

for N and P Management and Farm Performance.” Agriculture Ecosystems and
Environment 92, no. 2/3(2002): 283-296.

Ostrom, E, L. Schroeder, and S. Wynne. Institutional Incentives and Sustainable
Development: Infrastructure Policies in Perspective. Boulder, CO: WestvieW'Press.
1993.

Palmer, K., W.E. Oates, and RR. Portney. “Tightening Environmental Standards: The

Beneﬁt-Cost or No-Cost Paradigm?” Journal of Economic Perspectives 9(Fall 1995):
119-132.

Parker, D. “Controlling Agricultural Nonpoint Water Pollution: Costs of Implementing
the Maryland Water Quality Improvement Act of 1998.” Agricultural Economics
24(December 2000): 23-31.

Porter, ME, and C. van der Linde. “Toward a New Conception of the Environment-
Competitiveness Relationship.” Journal of Economic Perspectives 9(Fall 1995): 97-118.

Portney, RR. and RN. Stavins. “Introduction.” in P. R. Portney and R. N. Stavins (eds.)
Public Policies for Environmental Protection. Washington DC: Resources for the
Future. 2000.

Ruhl, J .B. “Farms, Their Environmental Harms, and Environmental Law.” Ecology Law
Quarterly. 27(2): 263-349. 2000.

158

 

 

Russell, CS. and RT. Powell. “Choosing Environmental Policy Tools Theoretical
Cautions and Practical Considerations.” in Marie L. Livingston (ed.) Environmental
Policy for Economies in Transition: Lessons Learned and Future Considerations.
Proceedings of the Resource Policy Consortium Symposium. Washington DC. May 20-
21, 1996.

Russell, C.S. Applying Economics to the Environment. New York: Oxford University
Press, 2001.

Segerson, K. “Flexible Incentives: A Unifying Framework for Policy Analysis.” in F.
Casey, A. Schmitz, S. Swinton, and D. Zilberman (eds.) Flexible Incentives for the
Adoption of Environmental Technologies in Agriculture. Norwell, MA: Kluwer
Academic Publishers. 1999a.

Segerson, K. “Mandatory Versus Voluntary Approaches to Food Safety.” Agribusiness.
15(1): 53-70. 1999b.

Schmitz, A., Boggess, W.G., and Tefertiller, K. “Regulations: Evidence from the Florida
Dairy Industry.” American Journal of Agricultural Economics. 77(5): 1166-71. 1995.

Shortle, J .S. and J.W. Dunn. “The Relative Efﬁciency of Agricultural Source Water

Pollution Control Policies.” American Journal of Agricultural Economics. 68(August
1986): 668-677.

Shortle, J .S., R.D. Horan, and D.G. Abler. “Research Issues in Nonpoint Pollution
Control.” Environmental and Resource Economics. 11(3-4): 571-585. 1998.

Simon, H.A. “Rationality as a Process and as a Product of Thought.” American Economic
Review 68(March 1978): 1-15.

Sims, T. Soil Testing for Phosphorus: Environmental Uses and Implications. Bulletin No.
389, SERA-IEG 17. University of Delaware, Newark, DE. 1998.

Srivastava, L., 8.8. Batie, and RE. Norris. The Porter Hypothesis, Property Rights, and

Innovation Oﬂsets: .T he Case of Southwest Michigan Pork Producers. Selected paper for
the American Agricultural Economics Association (AAEA) Annual Meeting, Nashville,

TN, August 1999.

Stavins, R.N. “Market-Based Environmental Policies.” in P. R. Portney and R. N. Stavins

(eds.) Public Policies for Environmental Protection. Washington DC: Resources for the
Future. 2000.

Stephenson, K., L. Shabman, L.L. Geyer. “Toward an Effective Watershed-Based
Efﬂuent Allowance Trading System: Identifying the Statutory and Regulatory Barriers to
Implementation.” The Environmental Lawyer. 5(3): 775-815. 1999.

Thompson, D.B. “Beyond Beneﬁt-Cost Analysis: Institutional Transaction Costs and the
Regulation of Water Quality.” Natural Resources Journal 39(October 1999): 517-541.

159

Thurow, AP, and J. Holt. “Induced Policy Innovation: Environmental Compliance
Requirements for Dairies in Texas and Florida.” Journal of Agricultural and Applied
Economics 29(July 1997): 17-36.

Tirole, J. The Theory of Industrial Organization. Boston: MIT Press, 1995.

Williamson, O.E. “The Modern Corporation: Origins, Evolution, Attributes.” Journal of
Economic Literature 19(December 1981): 1537-68.

Williamson, 0E. The Economic Institutions of Capitalism. New York: Free Press, 1985.
U.S. Department of Agriculture. Census of Agriculture. Washington DC. 1997.

U.S. Environmental Protection Agency. Ofﬁce of Water, Ofﬁce of Wetlands, Oceans and
Watersheds. Nonpoint Source Guidance. Washington, DC. December 1987.

U.S. Environmental Protection Agency. National Water Quality Inventory Report to
Congress 1998. Washington, DC: National Academy Press, 1998.

U.S. Environmental Protection Agency. Ofﬁce of Wastewater Management. NPDES
Frequently Asked Questions. Retrieved on February 18, 2003 from the World Wide Web:
httD://cfpub.epa.gov/npdes/allfaqs.cfm?program_id=0. 2002.

U.S. Environmental Protection Agency. Ofﬁce of Water. Overview of Current Total
Maximum Daily Load - T MDL - Program and Regulations. Retrieved on March 18, 2003
from the World Wide Web: http://www.epa.gov/owow/tmdl/overviewfs.htrnl. March
2003.

U.S. General Accounting Ofﬁce (GAO). Water Quality: Key EPA and State Decisions
Limited by Inconsistent and Incomplete Data. GAO/RCED-00-54. Washington, DC.
2000.

Xepapadeas, A.P. “Environmental Policy under Imperfect Information: Incentives and
Moral Hazard.” Journal of Environmental Economics and Management 20(March 1991):
113-26.

160