Assessing the implementation of the RECLAIM cap-and-trade market for pollution: measurement issues in counterfactuals, goal attainment, and command-and-control alternatives

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

ASSESSING THE IMPLEMENTATION
OF THE
RECLAIM CAP-AND-TRADE MARKET FOR POLLUTION:
MEASUREMENT ISSUES IN COUNTERFACTUALS, GOAL ATTAINMENT, AND
COMMAND-AND-CONTROL ALTERNATIVES

by

Christopher J. Wright

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POLITICAL SCIENCE)

December 2008

Copyright 2008 Christopher J. Wright

ii
Dedication

To Jerry, Kathy, Emma, Levi, Jesse, & Leah Rose.
And to my lovely bride Kim. Half this work and all of my heart belongs to you.

iii

Acknowledgements

Without the help of numerous people this work would never have gone from start to
finish. Thanks to my first dissertation advisor Sheldon Kamieniecki who originally
suggested that studying RECLAIM might be worth my while. After Sheldon left USC
for greener pastures, Jeb Barnes was kind enough to step in and help see me through the
process. Without Jeb this work would have remained nothing more than a good idea
wrapped up in a few disjointed conference papers. Dan Mazmanian, who also served on
the committee, had particularly helpful insights. And big thanks to Ann Crigler, who
despite her busy schedule took the time to help out.

iv

Table of Contents

Dedication ii
Acknowledgements iii
List of Tables v
List of Figures vi
List of Abbreviations vii
Abstract viii

Chapter 1 INTRODUCTION AND BACKGROUND 1
Chapter 2 MEASURING POLICY & IMPLEMENTATION 29
Chapter 3 METHODS & RESEARCH DESIGN 58
Chapter 4 RELATIVE IMPACT OF RECLAIM 97
Chapter 5 EFFICIENCY AND INNOVATION 135
Chapter 6 DISCUSSION AND CONCLUSION 169
References 182

v

List of Tables

Table 1.1 Health Effects of Regulated Pollutants 8
Table 1.2 National Ambient Air Quality Standards 8
Table 1.3 CAC vs. CAT Expectations 22
Table 4.1 Annual NOx Emissions for Compliance Years 1994 - 2005 104
Table 4.2 Annual SOx Emissions for Compliance Years 1994 – 2005 104
Table 4.3 Net impact of RECLAIM on SOx emissions 1994 – 2005 109

Table 4.4 Net impact of RECLAIM on NOx emissions 1994 – 2005 109

Table 4.5 RECLAIM vs. Non-RECLAIM universe 122

Table 4.6 NOx Emissions for RECLAIM facilities 124
Table 4.7 NOx Emissions for Non-RECLAIM facilities 124
Table 4.8 SOx Emissions for RECLAIM facilities 124
Table 4.9 SOx Emissions for Non-RECLAIM facilities 124
Table 5.1 CAC & CAT expectations vs. RECLAIM findings 138

vi

List of Figures

Figure 4.1 RECLAIM SOx emissions vs. RTC supply 1989 – 2003 102
Figure 4.2 RECLAIM NOx emissions vs. RTC supply 1989 – 2003 102
Figure 4.3 Tons of NOx traded 1994 – 2004 with and without price 114
Figure 4.4 Tons of SOx Traded 1994 – 2004 with and without price 115
Figure 4.5 NOx RTC Supply and Demand 116
Figure 4.6 SOx RTC Supply and Demand 116
Figure 4.7 Gross NOx Emissions: RECLAIM vs. Non RECLAIM facilities 126
Figure 4.8 Gross SOx Emissions: RECLAIM vs. Non-RECLAIM facilities 127
Figure 4.9 Changes in SOx emissions for RECLAIM vs. Non-RECLAIM % 129
Figure 4.10 Changes in SOx Emission from 1993 baseline 129
Figure 4.11 Changes in NOx emissions for RECLAIM vs. Non-RECLAIM % 131
Figure 4.12 Change in NOx emission from 1993 baseline 132

vii

List of Abbreviations

AQMP Air Quality Management Plan

CAA Clean Air Act

CAC command-and-control

CAIR clean air interstate rule

CARB California Air Resource Board

EPA Environmental Protection Agency

NAAQS national air quality standards

NSPS new source performance standards

RECLAIM Regional Clean Air Incentives Market

SCAQMD Southern California Air Quality Management District

SIPs state implementation plans

viii

Abstract

The creation of cap-and-trade (CAT) markets for pollution has long been
suggested as an alternative to command-and-control (CAC) regulation. Economic theory
suggests that CAT markets could achieve similar environmental outcomes to CAC in a
more efficient manner. In addition, CAT holds the promise that they may deliver
environmental improvements above and beyond CAC. CAT should also spur innovation
in abatement technologies and processes. This dissertation uses the Regional Clean Air
Incentives Market (RECLAIM) as a critical case study of the effectiveness of cap-and-
trade programs. The main focus of this dissertation is in assessing the effects of the
RECLAIM program on emissions levels. The main contribution of this dissertation is
that it provides critical insight into measurement issues surrounding CAT markets. Most
previous assessment efforts have used the level of tradable permits to assess policy
effectiveness of emissions reductions. This dissertation argues that permit levels may not
truly represent a CAC counterfactual. A better measure might be to compare emissions
levels of regulated entities under CAT markets to those of CAC in a similarly situated
environment. Using this latter measure this study finds that the RECLAIM program did
not lower emissions levels below those found under CAC. Efficiency gains under
RECLAIM may also have been overstated. Further, RECLAIM probably did not produce
innovation. An implementation framework is adopted to explain why RECLAIM was
never a pure example of a CAT market. As the policy has matured it has deviated farther

ix
and farther from the CAT ideal. Better measurement standards and more in depth
assessments should produce better inferences on the extent to which CAT markets work
under real world conditions. As CAT markets are becoming widely diffused, usable
knowledge about their potential and limitations is critical.

1
Chapter 1: Introduction and Background

Overview

The creation of markets for pollution has long been suggested as an alternative to
command-and-control regulation. Economic theory suggests that these markets could
achieve similar environmental outcomes to traditional approaches at a far smaller price
tag. In addition, they hold the promise of they may deliver environmental improvements
above and beyond command-and-control; and that they should spur innovation in
abatement technologies and processes. Since large-scale use of market instruments was
not widespread until the early 1990s, it was not until recently that these predictions could
be tested. Empirical evidence now suggests that the majority of these markets tend to be
both more effective and efficient than their command-and-control counterparts (Ellerman
et al. 2000). A survey of the literature on these markets suggests that there is something
approaching a consensus that the conditions under which markets should be successful
are being met, and that they are even more efficient and effective than once predicted.
This dissertation uses the Regional Clean Air Incentives Market (RECLAIM) as a
case study of the effectiveness of cap-and-trade programs. The main contribution of this
dissertation is that it provides critical insight into measurement issues surrounding cap-
and-trade markets. In assessing RECLAIM’s effectiveness, this dissertation shows that
measuring program success is a much more complicated matter than is often realized. It
is not enough to simply assume a program is working as intended because it meets
statutorily defined goals, nor is it fair to conclude that that program has failed if goals are

2
not met. Measurements of cap-and-trade markets should be more meaningful when they
are comparisons between the market as actually implemented and command-and-control
as actually implemented.
The secondary contribution of this dissertation is that it provides insights into how
cap-and-trade markets operate in the real world. The RECLAIM program is a
substantially important example of cap-and-trade markets. This analysis of the
RECLAIM program finds that it is just as effective as command-and-control programs
administered by the same regulatory agency. This analysis shows that the program does
not deliver emissions reductions quite to the extent claimed by the regulatory agency
tasked with its implementation. Further, it finds that efficiency claims about RECLAIM
are probably overstated and that the program has probably not led to greater innovation.
An implementation framework is adopted to explain how cap-and-trade in theory was
translated into the working RECLAIM program. Environmental and political factors
meant that the RECLAIM program was never designed to reflect the ideal of cap-and-
trade. The program continues to deviates farther and farther from the ideal, possibly as a
result of political pressures influencing the implementing agency.
This dissertation assesses the extent to which cap-and-trade markets are as
effective as predicted and as discussed in the economic and policy literature. The
conditions found in the greater Los Angeles area that are regulated by the South Coast
Air Quality Management District (SCAQMD) meet the theoretical conditions necessary
for cap-and-trade to be effective. RECLAIM is the nation’s first, largest, and longest
running regional cap-and-trade program. In addition to meeting the basic heterogeneity

3
criteria that economic theory predicts should lead to better policy outcomes than
command-and-control, RECLAIM has, at times, been heralded as a program to be
emulated nationally (EPA 2002; OECD 2002).

CAT diffusion

Knowledge about how cap-and-trade markets operate is particularly pressing: they
are spreading rapidly. In 2005 the EPA adopted new rules which lay the foundation for
two new cap-and-trade markets. The first, the Clean Air Interstate Rule (CAIR), is very
similar to RECLAIM in that it targets the same two chemicals regulated by that program-
- acid rain causing emissions and smog precursors. Unlike RECLAIM, though, which
only covers the greater Los Angeles basin, CAIR covers the entire eastern half of the
continental United States. The second is a new rule adopted by the EPA creating a cap-
and-trade system for mercury emissions. Another proposal, the ‘Blue Skies’ initiative
which would create another national cap-and-trade market, failed when it met
Congressional opposition, but the Bush Administration continues to support its passage.
In January 2005 the European Union implemented a cap-and-trade system to help
meet its greenhouse gas reduction obligations under the Kyoto protocols. If the U.S. is
ever to sign on to a modified version of Kyoto, it will almost certainly do so only with the
understanding that cap-and-trade markets will be central. Not only are cap-and-trade
markets spreading quickly, they are spreading worldwide.

4
How has RECLAIM performed? Answering that question is the central task of
this work. Theory predicts that cap-and-trade markets have the potential to lower
emissions, do so more efficiently, and produce innovation. In assessing RECLAIM
performance it is proposed that using the total number of tradable permits available as a
baseline measurement may not always be useful. This is particularly the case when
initial permit levels are unreasonably high and when program rules allow for mitigation
outside the cap-and-trade system. Because cap-and-trade programs are being rapidly
diffused, this dissertation follows the long tradition of policy studies which not only build
theory, but which also add usable knowledge that can help policy practitioners design
better policies (Lasswell 1971).
This dissertation uses the simple case study method. It uses the RECLAIM
program as a critical case of cap-and-trade. It compares program goals with program
performance, measured in a number of different ways. One important measure is to look
at what implementing agencies thought would have occurred under command-and-
control, and to how firms still under command-and-control in the RECLAIM region
actually behaved. This work shows that that expectations about what might have
occurred under different regulatory regimes are often misplaced.

RECLAIM as a critical case of cap-and-trade

Economic theory predicts that cap-and-trade markets should produce greater
benefits than command-and-control under two conditions: 1) there are large numbers

5
market participants; 2) regulatory targets are heterogeneous. The RECLAIM universe is
relatively large compared to other cap-and-trade programs. In any given year, over 400
facilities participate in the RECLAIM market. This is in contrast with Phase I of the
national Title IV SO2 program which only had 263 facilities (Ellerman et al 2000, 6).
Further, the Title IV program only included a single type of regulatory target—coal fired
electrical power plants—and are therefore homogenous. Targets for regulation in the
RECLAIM program are more diverse. The program has regulated a diverse set of actors
including energy companies, oil refineries, natural gas facilities, an after market auto
accessories manufacturer, an insulation producer, and even a printer (SCAQMD 1999).
Not only does RECLAIM, at first glance, have larger number of facilities than the
Title IV program, those facilities are far more diverse. RECLAIM meets both criteria
which economic theory predicts should lead to outcomes better than what is expected
under command and control. In fact, it meets these criteria better than Phase I of the Title
IV program which Ellerman et al (2000) and others claim is an example of a successful
cap-and-trade market. In addition to this RECLAIM itself has been, at times, held up as
an example of a cap-and-trade market to be emulated (OECD 2002). Past assessments of
the program have proclaimed it relatively successful (Kaminiecki, Shafie, and Silvers
1999; Kerr, Anderson, and Jaksch 2000; Johnson and Pekelney 1996).
At first glance RECLAIM is a most likely case for success as predicted by
economic theory. In fact, it should be more efficient than the Title IV program since
marginal gains are more likely to be exploited when differences between abatement cost
structures are increased as facilities become more heterogeneous (Hahn and Noll 1982;

6
Tietenberg 1985). Economic incentives for finding new abatement technologies are also
increased with diversity, so RECLAIM should also produce greater innovation than the
Title IV market. If cap-and-trade should work anywhere, it should work in the
RECLAIM program. RECLAIM is therefore a critical case study in the way in which the
theory of cap-and-trade meets practice (Eckstein 1975; King, Keohane, and Verba 1994).
Because it meets these criteria for success better than Phase I of Title IV it should
produce effective outcomes even more efficiently than the much heralded federal
program. This dissertation assesses the extent to which RECLAIM produced the
predicted results, with a more in depth focus on measuring the effectiveness of the
program at reducing emissions. But first we turn to a general discussion about clean air
policy in the United States.

Background: Clean Air Policy in the U.S.

In December 1952 a thick smoky fog enveloped London and led to the deaths of
at least 1,600 people. Similar events occurred in the Meuse Valley in Belgium in 1930
with 63 deaths and 6,000 illnesses and in 1948 in Donora, Pennsylvania, with 20 deaths
and 6,000 illnesses. These tragedies illustrate the easily identifiable adverse affects of air
pollution on human health. More often than not, though, it is difficult to establish a direct
causal relation between a specific air pollutant and a particular disease or death.
However, numerous studies suggest that such a link exists. For instance, strong evidence
suggests that pollution from the burning of fossil fuels leads to the premature death of

7
between 30,000 and 35,000 Americans each year. The elderly, infants, children and
adults with respiratory problems are particularly susceptible (Smith 2004, 85). Air
pollution is a real problem with real consequences to human health and quality of life.
In 1970 America began the first truly national effort to make the air safe for its
citizens to breath. The Clean Air Act (CAA) was one manifestation of the dramatic shift
that occurred in the way Americans thought about the relationship between humans and
the environment in the late 1960s. While human made air pollution has been around at
least as long as the campfire, the CAA encapsulated the vision of a society that no longer
saw pollution as something unpreventable or tolerable. The heart of this law was the
requirement that the newly established Environmental Protection Agency (EPA) set
national air quality standards (NAAQS) for six pollutants (Bryner 1995). These
pollutants are thought to be the most dangerous to human health with a variety of impacts
ranging from eye and throat irritation to pulmonary arrest (see table 1.1). Acid rain was
not a policy issue at the time of the writing of the 1970 CAA, although it did regulate S02
by imposing New Source Performance Standards (NSPS) which applied only to new
power plants (Ellerman et al. 2000, 13-14).

8
Table 1.1 Health Effects of Regulated Pollutants
Criteria Pollutant Health concerns
Ozone Respiratory tract problems, such as difficult breathing and reduced lung function;
asthma; eye irritation; nasal congestion; reduced resistance to infection;
premature aging of lung tissue; possible link to asthma; possible premature aging
of lung tissue

Particulate matter Eye and throat irritation; bronchitis; lung damage; impaired vision

Carbon monoxide Impaired ability of blood to carry oxygen; effects on cardiovascular, nervous, and
pulmonary systems

Sulfur dioxide Respiratory tract problems; permanent harm to lung tissue

Lead Retardation and brain damage, especially in children

Nitrogen dioxide Respiratory illness and lung damage

Source: Adapted from Rosenbaum (2002: 190) and Bryner (1995: 53)

Table 1.2 National Ambient Air Quality Standards (1990 revisions)
Pollution Standard Value
a
Standard type
b

Carbon Monoxide (CO)
8-hour average
1-hour average

9 ppm
35 ppm

10 mg/m³
40 mg/m³

Primary
Primary
Nitrogen Dioxide (NO2)
Annual arithmetic mean

0.053 ppm

100 μg/m³

Primary and Secondary
Ozone (O3)
1-hour average

0.12 ppm

235 μg/m³

Primary and Secondary
Lead (Pb)
Quarterly average

1.5 μg/m³

Primary and Secondary
Particlate <10 micrometers
(PM-10)
Annual arithmetic mean
24-hour average

50 μg/m³
150 μg/m³

Primary and Secondary
Primary and Secondary
Sulfer Dioxide (SO2)
Annual arithmetic mean
24-hour average
3-hour average

0.03 ppm
0.14 ppm
0.50 ppm

80 μg/m³
365 μg/m³
1,300 μg/m³

Primary
Primary
Secondary
Source: Rosenbaum (2002: 182)
a
Units of measure for the standards are parts per million (ppm), milligrams per cubic meter of air (mg/m³)
and micrograms per cubic meter of air ( μg/ m³).
b
The Clean Air Act established two types of national air quality standards: primary standards set limits to
protect public health, including the health of “sensitive” populations such as asthmatics, children, and the
elderly; secondary standards set limits to protect public welfare, including protection against decreased
visibility and damage to animals, crops, vegetation, and buildings.

9
States would be charged with enforcing these standards and were required to file
State Implementation Plans (SIPs) showing how standards were to be met. The states
were then expected to meet these standards within an aggressive timetable. In California,
responsibility for the creation of the SIP was delegated to the California Air Resource
Board (CARB). The federal CAA act was augmented in California by a state law
requiring similar results. CARB further delegates responsibility to regional air quality
boards which must then formulate an Air Quality Management Plan (AQMP), or a
regional plan on how responsibilities laid out in the SIP will be fulfilled. A major
responsibility of the regional boards is to control mobile sources of emissions.
The original act required that primary air quality standards be met no later than
December 31, 1975. This deadline has been pushed back time and again as states have
resisted the timetable and find it difficult to meet the rigid standards set by the EPA. The
CAA was amended in 1977 and again in 1990. The latest version of the act now expects
compliance in California no later than the year 2010, thirty-five years after it was
originally required. However, it is doubtful that even this goal can be met. In the late
1990s more than 46 million Americans, mostly in urban areas, lived in counties where
pollution levels exceed at least one national air quality standard (Rosenbaum 2002). The
Los Angeles basin is the most polluted region in the country and the only area with air
quality listed as “severe” by the EPA. The California Clean Air Act also requires
reductions in pollution, but further specifies that the Los Angeles area was to reduce air
pollution between 1990 and 2000 by 50% -- a 5% reduction in emissions each year for 10
years. Thus local AQMDs are required by both state and federal law to reduce air

10
pollution in their respective districts or face severe penalties, including the potential loss
of billions of dollars in highway funds (Thompson 2000).
As Ackerman and Hassler (1981) have illustrated in their in-depth study, the 1977
amendments to the federal CAA dramatically altered the choices available power
producers to reduce S02 emissions. The act required all major new sources of S02 to
install flue gas desulfurization facilities, generally referred to as “scrubbers”. The overall
impact was to put Western power-producers, and to a lesser extent those in the Northeast,
at a competitive disadvantage because they were essentially required to clean-up
emissions that were already “clean” because they burned low-sulfur coal. (Ellerman et al.
200). Because scrubbers were only required on new sources of emissions, Ackerman and
Hassler (1981) argue that this created a perverse incentive to building newer and cleaner
power plants. When the CAA was amended once again in 1990, acid rain was a primary
concern. However, unlike many of the other provisions of the CAA, the acid rain
reduction program is administered nationally.

Command-and-control vs. cap-and-trade

The first two versions of the CAA can be said to have followed a “command-and-
control” model of regulation. Command-and-control is characterized by direct
government intervention in the processes of industrial production. Under command-and-
control, government regulatory bodies not only set ultimate standards of air pollution
which are to be met, but also prescribe the ways in which these standards will be met.

11
Command-and-control requires that regulators know a great deal about the processes
from which air pollution is a byproduct.
Regulators must then decide amongst various control mechanisms which could
lower pollution levels. For instance, regulators could require that production processes
be altered to lower overall emissions before they reach the environment or they could
require the installation of technology at the end of production processes. After choosing
the method by which pollution will be reduced, command-and-control then requires
vigorous enforcement and constant readjustment in light of newer technological
advances. The complex nature of the process by which command-and-control regulations
are formulated, implemented, and enforced is one reason why it has been so difficult to
meet the aggressive agenda laid out in the CAA. The authors of the original act had
assumed it would be a relatively easy task for regulators to develop cost effective
abatement strategies. This has not proven to be the case.
The 1990 amendments to the CAA represent a dramatic shift away from a
command-and-control toward a model of regulation based on economic incentives and
market-like instruments. Although the EPA had experimented with market incentives in
the previous two decades, no large-scale program of the size proposed in 1990 had ever
been implemented. Title IV of the 1990 CAA amendments set up a system of tradable
permits which would be used to bring down levels of sulphur-dioxide (SO2), a major
contributor to acid rain. This cap-and-trade system is unlike command-and-control in a
number of ways. Most important is the fact that while it still sets overall pollution goals,
it leaves the decision of how to reach those goals to individual companies. Responsibility

12
for implementation at the firm level is shifted from government regulators to the private
sector. Government regulators are then no longer responsible for determining what the
best available technology is and what controls will be used, allowing a great deal of
discretion to the individual firm.
Economic models predict that this form of regulation can lead to similar amounts
of pollution reduction in a much more efficient manner than could be expected under
command-and-control (Hahn & Noll 1982). There is also the potential for greater
emissions reductions than would have been found under command-and-control and a
greater likelihood to spur innovation. Empirical evidence suggests that the acid rain
program has had the desired results of greater economic efficiency over command-and-
control (Ellerman et al. 2000). The Title IV program is touted as a great improvement
over other forms of regulation and is held up as a model to be emulated on a global scale
(OECD 2002).
In 1993 the SCAQMD, which has regulatory authority over the greater Los
Angeles basin, revised its AQMP to include a program modeled after Title IV.
RECLAIM is an effort to reduce air pollution in the Los Angeles basin in a manner that is
more efficient than the command-and-control alternative (SCAQMD 1996). Because the
program supersedes existing command-and-control regulations, it also intends to reduce
pollution in a manner that is at least as effective. Ten years after the policy’s initial
implementation, RECLAIM is also touted as a model for urban areas around the globe to
reduce air pollution (OECD 2002). Unlike the acid rain program, though, the extent to
which RECLAIM is actually producing the desired results is open to much more

13
question. Developments at the beginning of this decade have suggested that the
RECLAIM program may not have worked quite as intended, although the issue is far
from settled. Some have suggested that RECLAIM is less effective than the command-
and-control alternative, and price spikes in RECLAIM pollution credits during
California’s energy crisis have raised the possibility that efficiency gains have also been
overstated. Further, some environmentalists have accused programs such as RECLAIM
and the Title IV acid rain program of being inequitable and undemocratic.

From command-and-control to markets

Of all the plausible explanations about causes of environmental degradation, the
single most common reason cited is that markets fail to capture the true social cost of the
production of commodities. Following the lead of Pigou (1920) and Coase (1960), it is
argued that the price of any one good may not reflect certain costs that are external to the
production process sending signals that distort market demand for an overproduction of
these goods. These “externalities,” many have argued, are the primary reason why the
free-market, left to its own devices, cannot solve environmental problems (Kraft 2001).
Air and water pollution are the classic examples of externalities in the context of the
physical environment. As long as access to pollute air or water is free, manufacturers
will do so, and the price of their goods will not reflect the true social cost of the
production process imposed on the public.

14
With this in mind, solutions to environmental problems have traditionally
eschewed the market in favor of top-down policy prescriptions by regulatory agencies.
The market was seen as the problem, not the solution. Mazmanian and Kraft (1999)
have created a useful typology to help understand major environmental policies in the
context of the times they were created. The first major epoch of environmental policies
in the United States can be said to have followed a “command-and-control” model of
regulation (Mazmanian and Kraft 1999). For instance the Clean Air Amendments of
1970 charged the newly formed Environmental Protection Agency with coming up with
standards for air pollution based on a set of technology-based criteria. Best available
technology would be required on all new sources of pollution. New power plants, for
example, were required to install state of the art scrubbers to reduce certain types of
environmental contaminants.
Criticisms of command-and-control take many forms. First, while early gains
made in environmental quality were significant, each new dollar spent on regulation in
the form of command-and-control seemed to produce less and less actual improvement in
environmental quality (Dobra and Wendel 1999). For instance, Hahn and Hird (1991)
estimated that by 1990 the full cost of environmental protection was approximately
$2000 per U.S. citizen. Goklany (1999) has gone so far as to suggest that most of this
improvement, coming at such great cost, is the result of efficiency improvements, the
process of deindustrialization, and local regulatory efforts -- none of which are related to
federal regulation.

15
A second criticism has been that command-and-control regulations only provide
for some minimum to be met but provide no real incentives for minimums to be exceeded
(Dobra and Wendel 1999). Firms are simply required to follow set procedures without
considering new technological breakthroughs that may lead to further reductions in
pollution. In fact, under command-and-control, there are often difficult regulatory
barriers to overcome when changing from one technological fix to another. Command-
and-control also stifles innovation because of its one-size fits all nature.
An interesting example of the regulatory burden of command-and-control is
illustrative. In 2003 the County of Los Angeles ordered a new diesel powered electricity
generator for the Avalon community, located on Catalina Island just off the California
Coast. Because the new unit was considerably more efficient than the unit it replaced, it
would have lowered the amount of emissions from the small power-production facility on
the island. However, because the unit was public property and had to be transported via
ship to the island, the County had to file environmental impact reports with the
SCAQMD. These impact reports were not about how much less pollution would be
produced by the new generator, rather the reports were required because the ship being
used to transport the cleaner unit might contribute to air-pollution. Eventually, the
SCAQMD approved the transfer of the newer unit because, in its estimation, the total
reduction in emissions justified the transaction. Thus, the simple act of updating
equipment—even equipment that might significantly improve the environment—requires
lengthy and costly bureaucratic pre-approval.

16
The last major criticism is that command-and-control treats diverse sources of
pollution as if they were alike (Dobra and Wendel 1999). While large marginal gains in
pollution abatement were derived by prescribing best available technologies to industries
that had homogeneous characteristics, it has been widely recognized that any remaining
gains in environmental quality would have to come from an increasingly diverse set of
polluting activities. As the diversity of the targets of regulation increases, it becomes
difficult for centralized actors to come up with more and more solutions to problems that
may be unique to individual polluters. The contextual nature of specific instances of
pollution begins to take on a greater degree of importance as sources of pollution become
more heterogeneous. For instance, what do coal-fired electrical power plants have in
common with manufacturers of after-market auto accessories? Both types of industries
are targets for regulation of Nitrous Oxides in Los Angeles, but may require different
technological fixes.
These criticisms have led many to conclude that command-and-control was either
never necessary or not the best regulatory tool (Anderson and Leal 2001; Goklany 1999).
However, when taken as a whole, the experience with command-and-control suggests
that at least two conditions are necessary for command-and-control to be the best option
for environmental regulation. First, polluters must be relatively homogeneous. Second,
regulators must be able to identify readily technological fixes at a low cost (Dobra and
Wendell 1999).
Recognizing this, to dismiss command-and-control as a wholesale failure would
be overlooking a great deal of historical evidence. Empirical studies consistently show

17
that significant gains in air quality were achieved following the 1970 Clean Air
Amendments (Portnoy and Stavins 2000). It would be difficult to ascribe all of these
gains, as does Goklany (1999), to local politics, mere happenstance, and an artifact of
historical luck. For those that subscribe to this view, the push for market incentives
represents a particular ideological orientation in American politics that believes that most
government programs are ineffective and inefficient and that government interference in
the private sector ought to be minimized.
Despite these objections, command-and-control has had the desired effect of
reducing air pollution. For instance, in Los Angeles much of the historical improvement
in ozone levels, the primary cause of concern in the region, occurred before any move
away from command-and-control (SCAQMD 2002). The great early strides made in
pollution reduction can, in part, be attributed to regulatory agencies targeting
homogeneous sets of polluters with similar technological fixes.
Further, in the context of the dramatic rise of environmental awareness following
Earth Day, many early attempts at pollution regulation was explicitly concerned only
with the effectiveness of the policy. Costs of meeting environmental standards were
intentionally left out of the policy implementation equation. For instance, the Clean Air
Amendments of 1970 required the formulation of air quality standards without regard to
cost or the present ability to meet those standards. “If control technologies were not
available to meet these standards, Congress expected them to be developed by fixed
deadlines” (Kraft 2001, 105). If effectiveness is the standard by which to measure these
regulations, then certainly command-and-control must receive some credit.

18
The context of environmental regulation changed greatly after the economic
slowdown of the late 1970s and the deficit crises of the 1980s. Budgetary restrictions and
the political changes brought by the Reagan revolution drastically slowed the rate of
legislation aimed at large-scale environmental improvement (Kraft and Vig 2000).
Further, the aggressive goals set by the Clean Air Act, and other pieces of environmental
legislation of the 1970s, seemed as elusive as they had 10 years earlier. The environment
was becoming cleaner, but each new increment of environmental improvement was
becoming increasingly costly and was slower in coming. It seemed that the conditions
for which command-and-control represented the best policy option were disappearing.
New targets for regulation were becoming heterogeneous making similar technological
fixes difficult as a matter of policy prescription.
Just as the effectiveness of command-and-control were being challenged by new
historical circumstances, so too were the goals of environmental policies. A general
skepticism of the government’s role in society combined with an era of limited budgets
made efficiency a paramount goal. There was a perceived need for increasing
effectiveness and efficiency in policy implementation (Portnoy 1990; Savas 1987). The
argument was that, “government, due to its large size, lack of incentives, and
vulnerability to the influence of outside forces, is not in a position to provide services at
the least cost and greatest benefit on its own” (Kamieniecki, Shafie, and Silvers 1999,
109). No longer would a policy be judged simply by changes in environmental
conditions, but rather cost effectiveness and economic impacts would become part of the
equation (Kraft 2001).

19
Harnessing the self-interested behavior of individuals toward socially desirable
goals was not a new idea in the late 1980s, but the move towards markets during this
period does suggest that something fundamental had changed. These ideas had been
around for a while, but like many ideas, the time had not yet come for them to be placed
on the public agenda. Using the framework of Kingdon (1995), we might say that
streams of problems, policy, and politics opened a window of opportunity for policy
entrepreneurs to put market instruments on the public policy agenda. From the
perspective of Sabatier and Jenkins-Smith (1993), it could be argued that competing
“advocacy coalitions” had learned through iterative policy debates, transforming the very
nature of the policy question. Perhaps the transaction costs associated with a new set of
policy problems just became too high for command-and-control to make sense and
business interest pressure on lawmakers elicited more efficient institutional arrangements
(Weber 1998). Whichever is the most likely explanation, it remains clear that something
happened in the late 1980s which led to a widespread embracing of market mechanisms.
Command-and-control is now recognized as only one option in a larger policy
toolbox available for addressing market failures. In a classic essay on why environmental
problems are so often associated with common-pool resources like air or water, Garret
Hardin (1960) suggested that one solution to resource depletion was for government to
directly intervene. But Hardin, along with a number of neoclassical economists, also
noted that the assignment of property rights to resources that previously were defined by
open access was also a policy option. For instance, if the goal was to stop overgrazing of
cattle on the community commons, one way of reaching that goal was to simply regulate

20
how many cattle could graze on it – an example of command-and-control. But another
option, though, was to privatize the commons—a market solution. The owner of the
property now has a long-term interest in not depleting the resource on which he is
financially dependent. It was not the market per se that led to adverse environmental
effects; rather, a lack of defined property rights was to blame (Ostrom 1990).
Following the logic of Pigou (1920) and Coase (1960), economists argue that
what is needed is a way of internalizing the costs of environmental harms to the
companies that produced those harms. This could be accomplished through the use of
such instruments as taxes on pollution, deposit-refund systems, or tradable permits
(Portnoy and Stavins 2000). While direct taxes on pollution are more common in
Europe, tradable permits have been the market based policy instrument of choice in the
US. Unlike command-and-control, these systems do not prescribe the way in which
pollution reduction will be met, but rather leave strategies for abatement to the individual
firm. It is widely argued that market incentives will meet the same goals of effectiveness
as those under command-and-control in a much more efficient manner. Further, markets
are likely to result in the development of new and effective pollution control
technologies. Lastly, it is argued that markets are easier for the public to understand, and
therefore easier to legislate (Rosenbaum 2002). It would be difficult to imagine
objections to market mechanisms if, all things being equal, they delivered the same
outcomes in a cheaper and less intrusive way than command-and-control.
It is no wonder that interest in market-based instruments for environmental
protection has greatly increased since the mid-1980s (Portnoy and Stavins 2000). In

21
1989 the federal government set up a tradable permit system as part of an effort to meet
obligations under the Montreal Protocol. Soon afterwards, the EPA began to allow a
limited amount of credit trading for NOx and particulate emissions among heavy-duty
truck and bus engine manufacturers. That same year the most ambitious market based
program to date was passed. Under Title IV of the 1990 Clean Air Amendments a
tradable permit system was enacted with the intention of reducing sulfur dioxide
emissions by 10 million tons below 1980 levels. This market has been estimated to save
as much as one billion dollars annually. Regional and local markets followed. Twelve
northeastern states, and the District of Columbia, now participate in a nitrogen oxide
trading system, with compliance costs savings estimated at 40-47% savings over
command-and-control (Portnoy and Stavins 2000, 2). RECLAIM, in Southern
California, was one of the first local emissions trading program, but other states have
followed. Colorado, Georgia, Illinois, Louisiana, Michigan, and New York have all
established programs for the trading of nitrogen oxides and volatile organic compounds
permits (Bryner 1999).
From the above discussion, we see that command-and-control has a single goal:
effectiveness. This goal was expressed in legislation requiring strict reductions in
pollution regardless of the cost. Later, policymakers also included efficiency as a goal
for environmental policies. Cap-and-trade has been widely supported largely because
economic theory predicts that they are more efficient than the regulatory alternatives.
The incentives inherent in cap-and-trade programs also may lower emissions above and
beyond what we might find in command-and-control. Further, these incentives also

22
should lead entrepreneurs to come up with innovative abatement technologies. The goals
of command-and-control vs. cap-and-trade are summarized in Table 1.3 below. Chapter
five presents a similar table, but includes the findings of this dissertation of the extent to
which RECLAIM lived up to these goals.

Table 1.3 CAC vs. CAT expectations
Command-and-Control Cap-and-Trade
(Theory)
Goal Attainment Yes. Goals defined by
technology known to
regulators. Goals
adjusted up or down in
light of technological
improvement.
Yes + possibility of
exceeding goals.
Efficiency No. Yes. Efficiency gains
predicted based on
differences in marginal
costs of abatement
between facilities.
Innovation No. Reliance on
regulators to find best
available technology.
Yes. Technology
forcing as firms try to
find lowest cost way to
achieve goals inspiring
entrepreneurial activity.

A major shift occurred sometime after 1988 in which markets became more
favored than command-and-control as a method for certain environmental goals to be
achieved. During this time, the major goals of environmental policy moved towards both
effectiveness and efficiency. Markets became embraced as the favored implementation
strategy by many in the environmental movement, among business interests, and among
regulatory agencies. Further, this shift has occurred not only in the US, but also in

23
Europe, and is set to be at the center of future international agreements on limiting
greenhouse emissions (Helm 2000). With the age of many of these programs now past
the ten-year mark, and the move toward this form of regulation accelerating, it seems
appropriate to begin to ask the extent to which these programs are working as they were
intended.
Colby (2000), in a review of three cap-and-trade markets found that there is
considerable difference between optimistic ex ante economic models and actual costs and
benefits associated in creating markets where none previously existed. While she does
suggest that the move towards markets is a good one given the problems associated with
previous regulatory methods, she urges that more realistic assessment of these programs
need to be made. DeLeon (2001) notes one reason that implementation studies were not
done prior to the early 1970’s was that it was assumed that policies were being carried
out in expected ways. Are we making the same mistake? Are we simply assuming that
market oriented regulation is working as expected?
The goal of this dissertation is to assess the extent to which a single cap-and-trade
market, RECLAIM, is living up to a number of predicted expectations. It finds that the
RECLAIM is just as effective at lowering emissions as the command-and-control
alternative, but no more. Economic theory predicts that an efficiently run cap-and-trade
market might go beyond the minimum required in command-and-control. This is not the
case in RECLAIM. Further, this study finds that RECLAIM is probably much less
efficient than previous studies indicate. This has important implications for
understanding why minimal effectiveness expectations were not exceeded. Further,

24
RECLAIM probably has not led to innovation, also a causal factor in why emissions
were not reduced beyond the minimum. An implementation framework is adopted to
explain how cap-and-trade in theory is translated into a working program. The program,
as implemented, has never been a reflection of the ideal and over the years it has
undergone substantial changes which help explain why it is not as effective or efficient as
theory predicts.

Plan of the dissertation

Chapter Two
This chapter first reviews the implementation literature. It emphasizes how
markets are more than a policy tool, but are a distinct form of implementation.
Underlying much of the implementation literature is a causal story in which policies are
first made in Washington and then implemented at the street level by state or federal
bureaucrats. Much of this literature emphasizes the ways in which those at the bottom of
the implementation chain thwart policy intentions. The literature suggests that carefully
crafted policies and commitment to policy intentions by implementing agents are key to
policy success.
Next, it explores the conditions under which cap-and-trade markets should
successfully deliver more efficient and effective outcomes than their command-and-
control alternatives, as proposed by prior theories. What is underemphasized in these
theories is the connection between the creation of a property rights under market forms of

25
regulation, and the role of bureaucrats in enforcing them. It is proposed that a necessary
condition for a successful cap-and-trade market is a commitment by bureaucrats to the
property rights created. Since price is the mechanism by which the market signals the
underproduction of a good, a rise in prices in a cap-and-trade market is not necessarily
evidence of a failed market. Interference by well-intentioned bureaucrats in a market
experiencing a rise in prices reduces the value of the property right created by it, is
inequitable to market participants who have lower marginal abatement costs, and will
lead to distrust between firms charged with day to day implementation tasks and
regulatory agents.

Chapter 3
Chapter three provides the methodological underpinnings of the analysis that
follows. The case study method is identified as the most appropriate for examining the
small universe of programs that are both cap-and-trade and have been running for a
sufficient length of time to make meaningful inferences. Various measures of policy
effectiveness are discussed. This study adopts multiple effectiveness measure, but it is
argued that the most meaningful of these is relative impact. The pros and cons of using
levels of tradable permits as a baseline counterfactual is explored. It is argued that a
better method for assessing the relative impact of a cap-and-trade program is to
empirically compare program performance with a sample of similarly situated targets
operating under command-and-control. Efficiency is also explored as a dependent

26
variable. However, in this study efficiency is also treated as an independent variable
since it is a causal driver in ensuring that market programs are effective.

Chapter 4
Chapter four examines one aspect of RECLAIM effectiveness, relative impact. It
uses public data generated by the SCAQMD and the CARB as the primary data sources.
It is argued that the internal measure used by the implementing agency is implicitly a
counterfactual of the predicted path under command-and-control. These data show the
level of emissions allowed under the RECLAIM program compared to emissions levels
as measured by the SCAQMD. Special emphasis is given to pollution levels and permit
pricing in the 1999-2000 trading years. These data show that RECLAIM initially
exceeded internal forecasts and expectations, but at times has been less effective than
initially predicted. Overall, though, using this measure RECLAIM has had significant
impact above that predicted under command-and-control.
A closer look at the counterfactual’s construction is then taken. It is argued that
the counterfactual does not truly represent levels of emissions that would have occurred
under command-and-control. Instead, it is proposed that a comparative case study is a
more accurate reflection of the way the program performed. A detailed quantitative
comparison between how those in RECLAIM performed vs. how those outside the
RECLAIM universe performed is undertaken. This analysis shows that the
counterfactual used by the implementing agency is not valid. Further, it shows that
RECLAIM is not nearly as effective as once thought, but not nearly as ineffective as

27
some claim. Variation in levels of pollution among regulated firms cannot be attributed
to participation in the RECLAIM program.

Chapter 5
Chapter five examines the extent to which the RECLAIM program is as efficient
as the economic literature would suggest and if it has produced innovation. RECLAIM is
examined to see if it meets the criteria for a number of conditions under which a cap-and-
trade market should be efficient. This study finds that RECLAIM meets many of these
criteria, but on other points the answer is vaguer. What is often overlooked by standard
economic models of cap-and-trade efficiency is commitment to the rules and to property
rights. On these two points RECLAIM falls short. Because of this the SCAQMD have
not let the price mechanism work the way it should. Thus the program is probably much
less efficient than previous studies have concluded. This is important because efficiency
is a causal driver of effectiveness, and therefore this helps explain why RECLAIM did
not produce emissions reductions beyond what would have been found under command-
and-control.
This chapter also finds that RECLAIM deviates significantly from the cap-and-
trade ideal. In its initial design and throughout the life of the program changes have been
introduced which benefit certain politically powerful actors at the expense of others. This
is another explanation for why RECLAIM did not perform as well as cap-and-trade
theory predicts: it isn’t a pure cap-and-trade program. That policy is changed during the
implementation process is consistent with much of the early implementation literature.

28
Chapter 6
This work concludes by suggesting that cap-and-trade markets are likely to
continue to be diffused on a national and international scale. Because of this, studies of
how these markets work in practice rather than in simply in theory should be particularly
useful. Those with a vested interest in seeing these policies diffused should also have an
interest in discovering what went right and what went wrong with RECLAIM. The
prospect for more effective, efficient, and innovative policy results than command-and-
control make them considerably desirable for those in the policy process seeking to
overcome political barriers to environmental improvement. If markets are to work
correctly, then the property rights created by them must be respected by administrators
fearing the short-term political repercussions of a temporary price spike. Those with an
interest in assessing the extent to which alternatives to command-and-control are
effective are urged to find the most appropriate baseline. When possible, that baseline
should be empirically based and compare similarly situated targets of regulation. It is not
enough to know the failings of one type of implementation strategy without considering
that the next most likely alternative might also have failings which would lead to similar
outcomes.

29
Chapter 2 Measuring Policy

Implementation Studies: In search of a dependent variable

The process approach to the study of public policy (e.g., Jones 1978) is an
improvement over the institutional approach of traditional political sciences because it
recognizes that policymaking involves multiple institutions. It also draws generalizations
about policymaking without exclusively focusing on the legislative branch, the
bureaucracy, or other institutions with unique characteristics that could color one’s
conclusions (Sabatier and Jenkins-Smith 1993). At the same time this approach
conceptually divides the policy process into discrete stages. The informal process begins
as issues are defined and agendas are set as public concerns are translated into the
concerns of policymakers (Cobb and Elder 1983). The formal policy process begins with
policy formulation and enactment. Implementation is normally conceived of as a stage
that occurs late in the process as lower level bureaucrats translate broad policy goals into
formal rules and processes.
Implementation studies seek to explain the link between programmatic goals and
programmatic achievements (Knaap and Kim 1998). This field of study can be traced to
concerns of scholars in the early 1970’s that many of the promises made by Great Society
programs were not being fulfilled (Derthick 1972; Pressman and Wildavsky 1973;
Bardach 1977). Pressman and Wildavsky’s Implementation (1973) and Bardach’s The
Implementation Game (1977) are the seminal classics of the first era of implementation

30
studies. The former studied a federal redevelopment program which targeted the
Oakland area, the latter reviewed earlier literature on implementation and added several
case studies.
Pressman and Wildavsky view implementation as establishing chains of causation
between initial conditions and future consequences. So to them implementation is the
ability to forge subsequent links in the causal chain so as to obtain the desired results.
Implementation rests on what they call “the complexity of joint action.” According to
Pressman and Wildavsky, participants may agree with the objectives of a policy and still
oppose the means of carrying them out. The means are sometimes incompatible with
other commitments, they may have a preference for other types of policy tools, they are
simultaneously committed to other projects, they might depend on others who lack a
sense of urgency, and they may also have different opinions about leadership roles
(Rampedi 2003). It may even be the case that those involved in the policy
implementation chain agree with both the goals and the means, but do not have the power
to affect change. Successful implementation, then, is conditioned on limiting the number
of players in the chain and structuring incentives so that institutions and individuals
targeted by a policy will cooperate.
Bardach offers little in the way of disagreement with the core theory of linkage
offered by Pressman and Wildavsky. His major contribution was in describing the
implementation process as a series of ‘games’. Policies were more likely to be successful
when they are carefully crafted to anticipate these games. Thus, it is up to elected leaders
or those otherwise involved in policy formulation to ensure that those lower down the

31
chain faithfully fulfill their intentions. A second and related recommendation from
Bardach is that once policies are implemented that the game would need to be ‘fixed’.
Thus, Bardach sees implementation as not only a process which ‘begins’ in the legislative
branch and ‘ends’ when street-level bureaucrats officially start a program, but as an
iterative process. Implementation is equally about how the policy gets carried into action
on a day to day basis.
What is interesting about this vision of implementation is that in the second
edition of Implementation, Wildavsky and new co-author G. Majone (1983) admit in an
added chapter that implementation is an interactive process. Hill and Hupe (2002, 45)
speculate that Majone, being Italian, had a very different view of the way policies ought
to be carried into action than Wildavsky and Pressman. The European public
administration literature on the way in which public servants should behave has always
stressed the need for flexibility in the process. Flexibility on the part of those charged
with enforcing rules is seen as the natural response of street-level bureaucrats dealing
with pressures unforeseen by those that had developed the original policies (Lipsky
1980). That policies are inevitably altered by those who carry them out is admitted by
all, what is contested is whether or not this is desirable or not.
To some in this early cohort of studies, implementation was “the missing link” in
understanding how the good intentions of policymakers become good policies (Hargrove
1975). However, this first generation of implementation studies was fairly pessimistic
about the degree to which a public policy can be effective (Calista 1994). Much of this
can be blamed on the particular cases they chose to study. But discovering why many

32
Great Society programs seemed to fail was an important research agenda. At the bottom
of these failures had to be something deeper than, “They failed because they were hard to
do.” They described how faithful fulfillment of the intentions of policymakers can get
confounded because implementation must create “joint actions” out of previously
unrelated actors and agencies. These agencies and actors are not always committed to
constructing common outcomes (Pressman and Wildavsky 1973). Discrepancies may
also appear in outcomes because bureaucrats doing the implementing may restrain
themselves from embracing policy intentions fully (Bullock 1981). As a result, the most
prevalent finding in this first generation of implementation research is that outcomes are
either disappointing or unwitting, largely because of the incompatible goals or values of
those assigned the task (Calista 1994).
A second generation of implementation studies had the more aggressive agenda of
finding a generalizable theory of implementation. Many of these analyses criticize the
case study method employed by earlier researchers and urge those who study
implementation to be more rigorous in the way they model the policy process
(Mazmanian and Sabatier 1983). They also tend to be more positive about the potential
effectiveness of implementation than first generation studies and offer several conditions
under which policies are most likely to succeed (Mazmanian and Sabatier1983; Lipsky
1980; Hjern and Hull 1983). However, this second generation produced two schools of
thought, each proposing alternative hypothesis as to the best way to achieve policy
success.

33
The “top down” approach is exemplified by Mazmanian and Sabatier (1983). It
emphasizes the need for clear goals and specific directives to be handed from policy
makers to policy implementers. Mazmanian and Sabatier (1983, xii) consider three types
of independent variables that may affect whether or not policies will be affective:
1. The tractability of the problem being addressed, for example, the availability
of a valid theory of the factors affecting a problem and the extent of change
required in target group behavior

2. A variety of ways in which a statute or other policy directive can constraining
the implementation process through the clarity of its directives and the
assignment of rights and responsibilities to various actors

3. The nonstatutory variables, for example, the attitudes and resources of various
sovereign and constituency groups, affecting policy implementation

The first, “the tractability of the problem” deals with the fact that some problems
may simply be harder to solve than others. The other two are the “ability of the statute to
structure implementation” and the “non-statutory variable affecting implementation”,
such as political support (Sabatier and Mazmanian 1980, 544). Because the second of
these factors is the one primarily in the control of policy-makers, their recommendations
focus on them. Those tasked with implementing public policies may have varying levels
of commitment to the program, but a well-structured statute can limit the ability to thwart
that policies’ intention by uncooperative agents or those targeted by the policy.
Flexibility in the chain is not overly emphasized.

34
The “bottom up” approach emphasizes the role of “street-level” bureaucrats in the
implementation process and suggests a greater role for them in policy formulation
(Lipsky 1980; Hjern and Hull 1983). Hjern and Hull (1983) go so far as to suggest that
earlier studies had missed the point entirely: the “policy” to be studied is not the one that
is handed down by policy-makers, rather the “policy” is in fact what is implemented.
Policies, under this definition, are not intentions but are actions, and actions are what are
carried out in fact and not simply what is intended. Implementation studies, then, should
not be about how to get street-level bureaucrats to faithfully follow the intentions of
policy-makers, but are about the outputs created by those agents. Policies cannot be said
to have ‘failed’ when one changes the very definition of what constitutes a ‘policy’.
Except to those interested in organizational theory, the core critique of the
bottom-up approach offers little insight. The very research question being asked by this
group of theorists is not the same as those asked by the ‘top-downers’ of both the first
and second generations of implementation studies. The two are not, properly speaking,
competing theories. The top-down approach is interested in the effectiveness and
efficiency of public policies, the bottom-up approach is interested in organizational
behavior. As originally conceived, organizational behavior was considered an
independent variable affecting the dependent variables of effectiveness or efficiency.
The actions of those charged with implementation have become the dependent variable to
the bottom-up group of theorists. While sharing an interest in how bureaucrats
implement public policies, that is where their similarities begin and end.

35
In spite of this there are important lessons to be learned from the bottom-up
school of thought. Those charged with implementation might actually need to be
consulted during the policy formulation stage since bureaucratic flexibility might be
important. Hill and Hupe (2002) cite the work of Barrett and Fudge (1981), who are
characterized as bottom-uppers, to find an important lesson from Lipsky (1980) and
Hjern and Hull (1983), “if implementation is seen as ‘getting something done’, then
performance rather than conformance is the main objective and compromise a means of
achieving it” (Hill and Hupe 2002, 56). As long as implementers are committed to
programmatic goals, then giving them flexibility to achieve those goals is an important
attribute of a successful policy.
Goggin and colleagues (1990) later proclaimed that a “third generation” of policy
studies was needed. They argue that their “communications model” brings together
“bottom-up” and “top-down” theories of the second generation and improves over the
case-study method of the first generation by making states the unit of analysis.
Variations in implementation, in this view, can be explained by differences at the state
level. Using states as the unit of analysis also has the advantage of a large enough sample
size as to be able to use more rigorous statistical methods in testing theory. This model
added important incites into the nature of American federalism, however it has not
received wide attention.
At about the same time as this “third generation”, others, even those who had
been involved in earlier implementation work, began to claim that dividing the policy
process into “stages” was at the heart of the problem in coming up with a general theory

36
of implementation . Paul Sabatier’s rejection is the most puzzling since he was
intimately involved in some of those earlier formulations and because his work was so
well received (Sabatier and Jenkins-Smith 1993). Dividing the policy process into stages
is an interesting heuristic device, they argue, but theoretically useless. They claim that
concepts such as “implementation” have, “outlived [their] usefulness as a guide to
research and teaching” (Sabatier and Jenkins-Smith 1993, 1). By abandoning the stages
approach to the policy process Sabatier and Jenkins-Smith advocated a changed research
agenda from that of effective public policy to “policy-change and learning.”
If the process is the dependent variable under scrutiny, what exactly is the
research question about if not the process itself? Such work has important implications to
sociologists, organizational theorists, political scientists, and others who study process for
its own sake, but if one’s research agenda is about effective public policy, then
abandoning the process stages framework is not the way to go. The “Advocacy Coalition
Framework”, as proposed by Sabatier and Jenkins-Smith is an elegant description of the
policy process, but not one which serves the traditional interest of policy scholars in
producing usable knowledge. As DeLeon and Kaufmanis (2000, 12) put it:
The study of public policy should not be restricted to an exclusive
playground for academic theorists, if it’s not allowed to ‘play in Peoria” – if
citizens are not permitted to understand and even assume a role in what is
happening – then one needs to wonder seriously as to what we are doing, why,
and for whom.

Others argue that the problem with studying implementation is that it is not quite
clear what is being studied. Helen Ingram (1997) notes that implementation studies are

37
not conceptually clear, and now advocates the study of policy design. Ingram’s focus,
though, has shifted from effective policies to understanding why certain policies are
adopted and why others are not, and how these affect target populations. Confusion over
exactly what implementation was and how to study it once led Peter deLeon (1998) to
proclaim, “the study of policy implementation has reached an intellectual dead end” (p.
313).
The first generation of implementation studies was pessimistic about the prospects
of policy success partially because of unrealistic expectations as to just what policy
“success” and “faithful implementation” meant. DeLeon (1998), Sabatier and Jenkins-
Smith (1993), and others suffer a similar disillusionment by their unrealistically high
hopes that a generalizable theory of implementation can be attained. The hope by
Goggin and colleagues’ (1990) to make implementation studies “more scientific” seems
also to have failed if one views a universally generalizable theory as the measure of that
kind of success.
Many have abandoned traditional implementation frameworks because of a
fundamental misunderstanding of the original purpose of the studies. When Pressman
and Wildavsky first set out to study the implementation process what vexed them was a
simple question: why was a particular policy not working? While it is true that they
offered several reasons why programs are generally not likely to succeed, their study was
not aimed at fulfilling Lester and Goggin’s (1998) hope that a “meta-theory” could be
developed. In fact, the rather lengthy subtitle of their work suggests the limited goals of
early implementation studies: “How great expectations in Washington are dashed in

38
Oakland; or, why it’s amazing that federal programs work at all this being a saga of the
Economic Development Administration as told by two sympathetic observers who seek
to build morals on a foundation of ruined hopes.” Implementation, to them, explained
why some policies succeeded and others failed. Some programs never have the desired
outcomes because they were never implemented, and others didn’t achieve the desired
outcomes because they were implemented poorly.
Using this early work as a guide for where the field should be going, we can be
more realistic as to the successes and failures of implementation research. While the
sheer number of articles and books with the word “implementation” dropped off
dramatically in the 1990s, this is by no means an indication that implementation studies
have been abandoned. The study of why some policies seem to work better than others is
the central problem in virtually all of policy analysis. For instance, what often passes for
a ‘program review’ is simply the studying of how a public policy is enacted on a day to
day basis; in other words how it is implemented. Contrary to what Hill and Hupe (2002)
claim, the move in the 1990s toward “governance” paradigm is not an abandonment of
implementation; it is simply the embracing by government of a different form of
implementation. The implementation of public policies continues, regardless of what we
may choose to call the mechanisms used for it.
If the biggest failure has been in the search for a meta-theory, perhaps the most
promising development can be found in studies that search for mid-level, contingency
models of implementation. DeLeon (2001), in his comprehensive criticisms of the field,
finds that “contingency theories” are a way of adapting to the obvious complexity of

39
implementation studies. That is, rather than making one shoe fit all, various authors have
suggested that different conditions might require different implementation strategies.
DeLeon, noting the contributions of authors such as Matland (1995), Ingram (1990), and
Scheberle (1997), says that “there is no single ‘best’ implementation strategy, that the
strategy is very much contextual in terms of what the contingencies surrounding the
policy issue are and how they can best be addressed in terms of implementation” (p. 5).
Such an approach answers the question of why a meta-theory of implementation has not
been forthcoming: such a theory cannot exist and be valid.
If we treat policy implementation strategies as contingent upon the specifics of the
problem being addressed and the contextual factors surrounding the immediate issue,
then we come much closer to the original intent of these studies -- to determine why it is
that some policies seem to fail and others seem to succeed. Without attempting to create
a meta-theory of implementation, we can begin to ask questions similar to the ones that
were asked by those who saw the policy successes and failures of the Great Society.
However, concluding that an attempt at creating a meta-theory is folly does not
necessarily lead to the conclusion that implementation studies should completely forego
its positivist orientation in favor of only measuring success and failure in terms of
democratic participation and learning as Schneider and Ingram suggest (1997).
A return to the original intent of implementation studies necessarily means
following Wildavsky and Pressman’s (1973) logic that “implementation, to us, means
just what Webster and Roget say it does: to carry out, accomplish, fulfill, produce,
complete” (p. xiii). Inasmuch as the stated goals of most policies are in changing some

40
condition other than the “state of democracy” or creating “institutions for learning,” it
seems reasonable to suggest that to study implementation we must usually consider
whether stated goals are achieved. Implementation is the connection between the goal
and the result. Whether or not these goals are formulated in Congress, in a state
legislature, by a city-council, or in the rule-making stage they are essential. Whether or
not those results are measured as bureaucratic processes or changes in some actual
condition, they are important.

Evaluation: measuring policy success

Implementation studies have almost invariably been connected to some evaluative
system whereby degrees of policy success can be measured. Criteria for success could
be thought of as imposed by the language of policies themselves (Mazamanian and
Sabatier 1983), by the promises of politicians and agency personnel (Pressman and
Wildavsky 1973), or by street-street level bureaucrats and target populations (Lipsky
1980; Schneider and Ingram 1997) However, the extent to which a policy is working can
be difficult to assess because of poorly worded, confusing, or contradictory language and
goals written by policy makers (Mazmanian and Sabatier 1983). Because policies are
often the result of compromises between competing interests, such ambiguity and
inconsistency may be purposeful (Sabatier and Jenkins Smith 1993). Competing interests
can use such ambiguity as a way to prolong the implementation process, politically

41
accountable policymakers can use ambiguous policy messages to pander to different
constituencies, and implementing agencies can use unclear policy goals as a way to
ensure bureaucratic survival and other institutional goals (Moe 1991).
Competing visions of policy intent and what is “good policy” may exist between
different agencies tasked with the implementation process and with different interest
groups active in policy formulation (Pressman and Wildavsky 1973). It would be
difficult to imagine a common set of agreed upon outcome indicators when policy is
targeted toward changing the behaviors of a heterogeneous set of populations with
competing visions of the “good” (Schneider and Ingram 1997). More often than not, this
dilemma has been resolved by imposing external standards of evaluation which are
explicitly normative. These normative standards almost always include some measures
of effectiveness and efficiency. To a lesser extent they have included such democratic
norms as equity and participation.
While effectiveness is quite rightly identified as an explicitly normative goal by
Schneider and Ingram (1997), some effectiveness goal is nearly always written into
statutory language. One of the problems with the post-positivist research agenda is that it
treats all goals as inevitably normative, thus freeing the researcher to impose her own
bias into the research design. If goals are contestable then why should the researchers’
own goals not be of equal value to those of electorally accountable legislators? By
imposing dependent variables other than those assigned in legislative intent, the
researcher may be providing a valuable service. That service, however, is one of

42
advocacy. One way to avoid this bias is to accept the policy goals as outlined in statutory
language.
Sometimes there may be conflicting goals written into statutory language; for
instance, the desire to drastically reduce air pollution in a short period of time, but to do it
in a way that has limited economic impact. However, when goals seem to be in conflict,
it is wise for the policy analyst to treat these goals as hierarchically ordered preferences.
For instance, since the 1980s most pieces of environmental legislation have included a
provision that cost-benefit analysis be included before rules are implemented. Thus,
effectiveness and efficiency are both written into the law. However, when understood
properly, these goals are not seen as competing, but rather complimentary. The first order
goal is to affect some environmental outcome (cleaner water, cleaner air, more of a given
population of a species of animal, more wilderness, etc.). Efficiency is a second order
goal. The first order goal of environmental effectiveness ought to be achieved, but in
such a way that is most efficient. The legislative goal of environmental policy is not to
run an efficient program, that is a general goal common to all public policies whether
stated in enabling legislation or not. Both efficiency and effectiveness may be treated as
dependent variables, but these variables need not be treated as if they were at cross-
purposes. Certainly, the fact that a policy may have multiple legislative goals ought not
give license to those evaluating it to disregard them as dependent variables altogether.
The first task of any implementation study is to asses the extent to which certain
standards are being met. Lynn’s (1987) idea of “backward mapping” might be the best

43
way to describe how implementation studies are usually done. For the most part these
studies have started with the premise that a policy has already failed or succeeded and
then worked backwards at finding out the weakest/strongest links in the implementation
chain. This study uses a similar approach. However, because this work takes seriously
the notion that policy “success” and “failure” are contestable terms, it seeks to evaluate
the extent to which cap-and-trade programs are really working as intended. It may be
that some success criteria are being met and others are not. It is only after we have fully
explored this question that we can go back and connect the dots between the intent of the
program and the extent to which it has been successful.

Market Failures vs. Government Failures

While using the tools of economic analysis is nothing new to public sector actors, the
realm of economics has been traditionally thought of as secondary. Efficiency concerns
are common to all public policy questions. Braden and Kim (1998) have noted that
instead of serving only as a tool to analyze those programs, economics are now being
used through markets as a key part of the implementation strategy. Dobra and Wendel
(1999) claim that using markets to forward environmental goals is such a dramatic policy
shift that big questions arise over the ability for bureaucracies to adapt. They raise three
main questions over market systems implementation:

44
1) ..how can and agency with minimal experience with trades use appropriate
caution and develop the system in phases?

2) …can the trade system assure that environmental standards will be maintained at
the level established under [command-and-control] regulation?

3) …how will the agency adapt to the new system? (p. 90)

Their second concern is central to this study. How do cap-and-trade markets fare
compared to the command-and-control alternative? While Dobra and Wendell raise the
possibility that market systems may not be as effective as command-and-control they find
little evidence that this is the case in the real world. Their main concern is not with the
effectiveness of cap-and-trade and other markets, but with efficiency. It should be noted,
though, that their efficiency concerns have less to do with markets per se, but with the
very real chance that bureaucracies lack the ability to create the mechanisms necessary
for markets to work as they should.
They do see the initial allocation of credits as a potential effectiveness problem.
Will the permits be allocated to current users or to the highest bidder? If given to initial
users, this puts them at a significant advantage and restricts competition. On the other
hand, auctioning to the highest bidders might shut down current operations. Who then
can buy the permits? Can environmental groups buy them? All of these questions have at
least some implications on the market program’s effectiveness (Dobra and Wendell 1999,
88). If, for instance, outside groups may buy pollution credits on the open market, then
there is the real potential that the market system would not only be as effective as
command-and-control, but even more so. This is because environmental groups might

45
then take those pollution credits and simply “retire” them. Each ton of pollution bought
by such a group would, under a system where the “cap” was at least as stringent as
command-and-control, mean one ton less of that pollutant was emitted.
Citing Posner (1992), Dobra and Wendell (1999) want to remind those in the
environmental movement skeptical of markets that many of those same criticisms can be
leveled at government intervention. In deciding whether government intervention is
appropriate, it is not enough to demonstrate that markets operate imperfectly;
governments also operate imperfectly. “What is necessary is a comparison between the
actual workings of the market and of government in a particular setting” (p. 84).
Too often an existing program is criticized without regard to the available options.
Such criticisms may be valid and may have the potential of pointing out much needed
room for improvement, but what is often overlooked is that the alternative options may
also be open to the same criticisms. The crux of Dobra and Wendell (1999) is that while
there may be problems with market mechanisms for environmental governance, similar
problems also exist for the alternatives. What does not exist in the command-and-control
alternative, though, is the potential for environmental effectiveness gains above the
minimum required by regulation. Thus, market mechanisms such as cap-and-trade are
generally a better choice.
The majority of studies done to date on cap-and-trade markets that have either
replaced or compliment traditional command-and-control have generally used the
programmatic “cap” as a baseline from which to measure success. As long as the total
amount of pollution emitted stays under the cap, the cap-and-trade program in question

46
will have, by definition, “succeeded”. Since the underlying assumption in these
programs is that the “cap” represents the total amount of pollution that would have been
allowed under command and control, staying at or below the “cap” is seen as evidence
that the program does not allow emissions to any greater extent than command-and-
control. By this measure nearly all of them come to the conclusion that the market in
question is as effective as command-and-control in delivering environmental
improvements, and then move on to measuring the market’s efficiency. When these
studies speak of “effectiveness”, they generally mean the effectiveness of the
implementing agency in creating an efficient market.
How efficient are these markets? Very efficient if the literature is to be believed.
Especially when compared to the heavy costs associated with command-and-control.
Stavins (2000) estimates that the cost-savings of the national Title IV S02 program
approaches $1 billion annually compared to the regulatory alternatives. Twenty years
after regulators, utilities, and environmental advocates first were introduced to air
emissions trading opportunities through the then experimental Emission Reduction Credit
programs of the 1970s, cautious exploration has evolved into a mature, productive
allowance market (Colby 2000, 10).
Cost savings have been so great, and the effectiveness of these programs so
seemingly self-evident, that Tietenberg (2000) claims another major benefit of these
programs: they are more politically feasible than further command-and-control
regulations. Indeed, the 1990 Amendments to the Clean Air Act were a major

47
breakthrough in a decade long political stalemate between environmental groups
demanding that coal burning power plants be cleaned up, an industry opposed to even
more regulatory interference in the day to day operations of facilities under their control,
and Republican administrations skeptical of the traditional bureaucratic approach to
regulation. If Tietenberg is correct, then markets may be even more effective than they
seem at first glance. If markets are more sailable than the alternative, then it seems likely
that any further command-and-control regulation may never have been adopted in the
first place. Or at least they may have taken some time longer to adopt than the market
alternative. If this is the case, then the “cap” in cap-and-trade does not really represent
the upper limit of emissions which would have occurred under an imagined command-
and-control. Since such a system may never have been adopted then any new limits set
would represent significant gains. The real alternative, then, is not an imagined set of
stringent rules but the present set of rules which would not be changed because of
political impasse.
This is not to say there have not been problems with these markets. Colby’s
(2000, 10) observations about the early years of the Title IV SO2 program affirms Dobra
and Wendell’s (1999) cautionary note that creating markets where none existed before
can be problematic. Even after trading rules become established, early trades on the
market still involve high transaction costs for several reasons. First, there remained
ambiguities to be clarified about the program and mechanisms for trade. A second reason
for high transaction costs was the lack of information on the market value of the tradable

48
right, a hurdle that was eventually overcome by public reporting of transactions. A third
reason that early transfers involve high costs were objections by those parties who were
opposed to creating a market in the first place. Would the market last, or would the
experiment come to a sudden stop because of a court ruling or shift in the political winds?
Each of these factors was eventually overcome in the SO2 and other cap-and-trade
programs, lowering transaction costs, and leading to a more efficient market. Even
though Colby is making observations about market efficiency, a cautionary lesson can be
learned from these observations about the effectiveness of a program as well. Drawing
conclusions about a program’s success or failure too early in the life of that program
comes at the risk of not anticipating how those participating in the program – whether
they be the implementing agencies themselves, targeted firms, or other market
participants such as environmental groups – will react and adapt to unforeseen
circumstances.
Glazer and Rothenberg (2001) make similar observations about the early
problems of the Title IV program and high transaction costs associated with the program
due to lack of experience and the political uncertainty surrounding what, to many
involved, seemed like an experiment which could, given a shift in the political winds,
come to an end. Credibility, to them, is key: “Although several economic constraints are
relevant for regulating firms, credibility is especially important for understanding what
government can and cannot accomplish with regulation” (p. 74). Is the likelihood of the
program’s long term political feasibility credible? And are those charged with
implementing the program have a sufficient commitment to it and to enforcing the rules.

49
When asked to make costly investments to meet regulations, firms may fear that a
government will later change its policies, or even that it has little intention of enforcing
the policy. Government officials may worry that interested parties will either doubt the
government’s resolve to punish firms for not making the requisite investments required to
bring them into compliance. They may skeptically view government’s willingness to
continue its policy and to provide the promised return on investments (p. 76).
The problem of credibility is not confined to market based programs, but are
endemic to all government regulation. But credibility becomes especially important
given that the creation of markets is also the creation of a property right (Dales 1967).
The right to pollute invested in a tradable permit is valuable given that a firm may sell
that permit on the open market. Programmatic changes, such as redefining who has the
right to buy or sell in the market, affects the value of the property right vested in a permit.
The major problem foreseen by Glazer and Rothenberg (2001, 86-88) in using
market mechanisms is not the misuse of markets by private actors, it is government.
Markets that are created by government can be destroyed by government. Will a market
system continue once it is adopted? And once adopted, how can consistency be
guaranteed?
A cap-and-trade system will work only if regulated polluters make investments to
reduce emissions. But firms will invest in emissions reductions only if they believe that
emissions will continue to be restricted. With such beliefs, investments will be made,
political support for regulation will increase, and an efficient market will be attained. By

50
contrast, the market will function poorly if firms believe government will change policy.
Investments in emission reducing equipment or process changes make little sense if the
property rights vested in an emissions credit is insecure. Support for continuing the
system will be weak, the market will not help to equalize emissions costs across the
spectrum of participants, few trades in permits and little investment will be realized.
Ultimately, with little outside political support and few trades, changing political fortunes
may lead the market to disappear. The culprit behind the imagined failed market?
Government agents uncommitted to maintaining the market or its rules. “Put another
way, can regulators be excluded from the regulatory market?” (Glazer and Rothenberg
2001, 87).
A necessary condition for environmental markets to work, then, must be a
fundamental commitment to maintaining and enforcing the rules. There is near
unanimous agreement on this point in the literature coming out of both economics and
out of policy studies. With the exception of Glazer and Rothenberg (2001), though, the
emphasis on government credibility has tended to focus on bureaucratic commitment or
ability to enforce the rules. The fear has largely been that agency commitment to police
firms, and that private actors would then take advantage of this lack of rule enforcement
leading to increases in pollution, lowered demand for tradable permits, and market
inefficiency.
For instance, Hahn & Noll (1982) claim that three conditions are necessary for
these types of programs to work: there must be an adequate number of participants in the
market, there must exist low transaction costs, and government must provide sufficient

51
monitoring and enforcement. Kamieniecki, Shafie, & Silvers (1999), in their review of
the RECLAIM program, claim that these three conditions have been met and that
therefore the market for pollution permits it created was efficient. They do give voice to
the concern that markets cannot produce desired outcomes if there is a perception that the
program which created the market will cease to exist. “Therefore, trading will collapse if
the government fails to maintain the program, because tradable permits have no value
outside a well-defined emissions market” (p. 114). But like most others they are more
concerned with how monitoring and enforcement might affect market efficiency. Thus,
they come to the conclusion that sufficient bureaucratic commitment exists on the part of
the SCAQMD because of the installation of real time emissions monitoring systems and
extensive yearly audits.

Property rights and efficiency as causal driver of effectiveness

Two things are either ignored or underemphasized in both the theoretical
literature on cap-and-trade markets and program reviews of them. First, bureaucratic
commitment is generally seen in terms of monitoring, enforcement, and sometimes as a
long term commitment to maintaining the market. With the exception of Glazer and
Rothenberg (2001), there is virtually no discussion of how a lack of commitment to
maintaining rules is key to maintaining the value of the property right vested in a tradable
permit. An implementing agency, for instance, may be committed to monitoring firms to
make sure they are not “cheating” and thus emitting more than they are permitted under

52
the program. The agency and its personnel may also embrace the values of the market
and be committed to its long term survival. But even when such a commitment is found
there may be underlying routines and actions which may undermine the very market they
are committed to maintaining. For instance, when an implementing agency perceives a
problem in a program which it is committed to, the normal process would be to attempt to
alter the rules in such a way as to improve them. While some adjustments may be
necessary for markets, especially in the early years as implementing agents begin to see
that real world contingencies are not always planned for, in the longer run rules changes
in a cap-and-trade market demonstrates a lack of commitment. Not a lack of
commitment to the market, nor a lack of commitment to enforcement, but a lack of
commitment to property rights.
Dales (1967) defines the ownership of property rights this way:
(1) a set of rights to use property in certain ways (and a set of negative rights or
prohibitions, that prevent its use in other ways); (2) a right to prevent others from
exercising those rights, or to set the terms on which others may exercise them; and (3) a
right to sell your property rights. (p. 173)

In a cap-and-trade market the tradable permit is the right to emit a given amount of
pollution. This permit meets all of Dale’s conditions of property. The right to emit
pollution allows firms to operate in ways they best see fit, as long as they do so under the
terms of the permit. The negative right associated with the permit is the cap it places on
pollution at the facility or firm level. But an important factor in determining the security
of this property right is the knowledge that those outside the system will not be able to
exercise the same right to pollute as those inside the system. This is usually expressed in

53
the literature on cap-and-trade as knowledge that there is not widespread cheating. But
what it also means is knowledge that the value vested in the tradable permit will not be
diluted through the addition of more and more participants. Conversely, it also means
that the value of the tradable permit will not be diluted through regulatory changes which
might limit the ability of a firm to sell.
Hix, Ruder and Sugarman (2000) note that the reason cap-and-trade markets for
Volatile Organic Compounds have not been as successful as other similar programs is
that the nature of this class of pollutants is such that strict regulatory oversight of each
transaction was necessary. This meant that implementing agents had the power to deny
trades and led to increased transaction costs and market inefficiencies. An inefficient
market means that firms are not sure what the clearing price of a tradable permit might
be. This leads to fewer investments will be made in pollution abatement technologies by
firms.
Changing the rules of who may or who may not participate in the market, or what
trades may or may not be done should have the same affect. The more implementing
agents insert themselves into the process the less sure market participants are of the value
of the property rights vested in the tradable permit. Will the market be flooded with new
permits given to new participants? Will the market dry up as regulations redefine who is
allowed to sell to whom? In either scenario, uncertainty about regulatory commitment to
property rights leads to market inefficiencies.
The second and related point is that the literature often underemphasizes the
causal relationship between market efficiency and effectiveness. Studies after study of

54
cap-and-trade markets proclaim that these markets have been effective, and then spend
quite a bit of time describing the extent to which these markets have been efficient. Both
are treated as important social values or programmatic goals. As I have already argued,
they should not be seen as competing goals but complimentary and hierarchical. Program
effectiveness being the primary goal –in the case of air pollution lowering of overall
emissions which would lead to improvement in human health – and program efficiency
being a secondary goal shared by all other public policies and usually prescribed in
statutory language. Both are also worthy dependent variables in implementation and
other policy studies.
However, in cap-and-trade and other market systems of regulation, market
efficiency is also the causal driver of programmatic effectiveness. An inefficient
government program is always problematic in that it wastes public resources. But it is
also possible to have an inefficient government program that also relatively effective at
producing valuable social goods. As we have seen, with few exceptions there is
widespread agreement that command-and-control has produced a great deal of
environmental improvement. But inefficiencies in command-and-control are so
widespread that pointing them out has been virtually a cottage industry for economists
and policy scholars for well over thirty years. Because the primary driver behind the
reduction of pollution in a cap-and-trade market is the firm’s willingness to invest in
technological or other improvements then it is impossible to imagine a scenario where
these markets could be inefficient and yet still effective. Property rights are only secure
inasmuch as they are protected by a system of law (Dales 1967), or in this case, by the

55
rules. When property rights are incomplete there is a disincentive to invest and pollution
is not abated.
To see why this would be, imagine a cap-and-trade system where rules are
constantly changed. We need not impute ill motives on behalf of regulators. Perhaps
they are making adjustments to the program because they believe such adjustments are
necessary in improving market efficiency. The constant reevaluation and adjustment of a
program has become routine for agency personnel because this is how implementing
agents are expected to act under command-and-control. Tinkering with the rules of a
program is just what they do. These regulators measure efficiency by monitoring the
clearing price of tradable permits on the open market. Thus, they believe low market
prices represent an efficient program. What would happen if these well intended
regulators with a lack of commitment to property rights were to set an artificial price cap
on tradable pollution permits? Firms who might otherwise invest in technology to abate
pollution will find little motivation in doing so now. There is no motivation for investing
in abatement for profit. If we assume these firms are in business to make money and that
their sole reason for existing isn’t to produce less pollution, then high prices for an
emissions permit are motivation for entrepreneurial innovation. Firms that can find ways
to emit less pollution in a less costly manner can sell excess credits on the open market
for profit.
Price is the mechanism by which markets allocate scarce resources. This was true
in Adam Smith’s day and remains true in all markets for all goods. Even when those
markets are artificially constructed and where the demand for goods in those markets

56
relies solely in the participant’s faith in agency commitment to the rules which created
those markets in the first place. When no such commitment exists, then market
participants cannot be sure that they will see a return on any investment made in
abatement technologies. If the primary effectiveness argument for cap-and-trade markets
is that they, unlike command-and-control, provide an incentive for firms to go beyond
minimum agency requirements, then removing that incentive should have a negative
effect on their potential to more effectively lower pollution.
Once an overall cap has been set in cap-and-trade and once permits are allocated,
then the primary role of regulators should be monitoring and enforcement. Any change
in rules that affects price also will affect the amount of pollution emitted. However
strong the temptation is for regulators to adjust and readjust market rules, any actions
they take will affect the program’s effectiveness. What may be appropriate agency
behavior in a command-and-control system might be disastrous under market forms of
regulation. Glazer and Rothenberg’s (2001, 87) question, “can regulators be excluded
from the regulatory market?” takes on added importance when viewing market efficiency
as the causal mechanism by which the program’s effectiveness is driven. If the
underlying causal theory is correct, then an inefficient market for pollution is also an
ineffective one.

57
Conclusion

What would cause regulatory agents from creating an efficient market? The
implementation literature can be quite illuminating on this front inasmuch as
implementation studies try to connect policy intents to policy achievements. One of the
most important lessons of early implementation studies is that government programs
might include conflicting goals representing the conflicting values of those involved in
the policymaking process. They also warn that during the implementation process
powerful interests may arise to try to thwart policy intent and therefore effectiveness.
Most importantly they show that a policy on paper is often far removed for how it
operates in practice.
In this chapter we have seen how implementation studies were first used to
explain the failures of Great Society programs. These studies help explain what happens
when programmatic intent is translated into programmatic practice. The economic
literature on cap-and-trade was then used to show the conditions under which these
programs should be most efficient and therefore most effective. In the next chapter we
see why RECLAIM, on its face, meets theses conditions and therefore economic theory
predicts that it should be very effective. Methodological and measurement issues in
assessing program performance are then addressed.

58

Chapter 3 Methods and Research Design

The central question asked in this study is to what extent does empirical evidence
show that cap-and-trade markets are as effective as the command-and-control regulations
they replace? This study uses the Los Angeles region’s RECLAIM program as a case of
how cap-and-trade systems perform under real world conditions. The program is
assessed in three ways. First, the program is assessed using quantitative data and
methods set by the implementing agency for internal evaluation. Because the baseline is
supposed to represent emissions levels which would have occurred had the original
command-and-control system been kept in place, this is used as a counterfactual case for
comparison. Second, the program is assessed using quantitative data representing how
much facilities actually emit in the same region and under the auspices of the same
implementing agency. Third, the program is assessed using quantitative and qualitative
data showing the conditions under which cap-and-trade programs ought to be efficient
and produce innovation. An implementation framework is adopted to explain the causal
drivers of the dependent effectiveness variable in the case of RECLAIM.

RECLAIM as a critical case of cap-and-trade

Proponents of cap-and-trade markets claim that they should be more efficient than
and just as effective as the more common regulatory regime, command-and-control

59
(Tietenberg 1985). In addition, an efficiently run cap-and-trade market offers incentives
for regulated firms to reduce emissions beyond minimum expectations. Cap-and-trade
therefore offers the prospect that it may be more effective than command-and-control.
Further, cap-and-trade also should produce innovation in abatement technologies and
processes above and beyond what would be found under command-and-control.
With these benefits in mind, cap-and-trade markets have become the tool of
choice for nearly all proposed new federal clean air regulations. Cap-and-trade markets
are also key to the European Union’s strategy for meeting its commitments under the
Kyoto protocol. If the U.S. is ever to join the global pact limiting greenhouse emissions,
it will doubtless only do so with the understanding that our commitments will also be met
with a cap-and-trade market. In anticipation to this, California has begun the process of
implementing its own greenhouse gas market. These markets are spreading fast.
Economic theory predicts that cap-and-trade markets should produce greater
benefits than command-and-control when there are large numbers of diverse regulatory
targets. At first glance, those facilities targeted by the RECLAIM program meet these
two conditions. The RECLAIM universe is relatively large. In any given year, over 400
facilities participate in the RECLAIM market. Phase I of the much heralded federally
administered Title IV SO2 program only included 263 facilities (Ellerman et al 2000).
Further, the Title IV program only included a single type of regulatory target: coal fired
electrical power plants. These plants have reduced SO2 emissions by either installing
“scrubbers” or by switching to “clean” coal—coal with a low sulpher content. Targets
for regulation under the Title IV program are homogenous. In contrast, targets for

60
regulation in the RECLAIM program are heterogeneous. The majority of emissions from
facilities in the program come from what Thompson (2000) calls “the big three”: energy
companies, oil refineries, and the local natural gas monopoly. However, the program has
also includes facilities such as an after market auto accessories manufacturer, an
insulation producer, and a printer (SCAQMD 1999).
Not only does RECLAIM, at first glance, have larger number of facilities than the
much heralded Title IV program, those facilities are far more diverse. RECLAIM meets
both criteria which economic theory predicts should lead to outcomes better than what is
expected under command and control: 1) large numbers of participants and; 2)
heterogeneous targets. In fact, it meets these criteria better than Phase I of the Title IV
program which Ellerman et al (2000) and others claim is an example of a successful cap-
and-trade market which serves as a model for cap-and-trade markets to come. But in
addition to this RECLAIM itself has been, at times, held up as an example of a cap-and-
trade market to be emulated (OECD 2002). Past assessments of the program have
proclaimed it successful (Kaminiecki, Shafie, and Silvers 1999; Kerr, Anderson, and
Jaksch 2000; Johnson and Pekelney 1996).
At first glance RECLAIM is a most likely case for success as predicted by
economic theory. In fact, it should be more efficient than the Title IV program since
marginal gains are more likely to be exploited when differences between abatement cost
structures are increased as facilities become more heterogeneous (Hahn and Noll 1982;
Tietenberg 1985). Economic incentives for finding new abatement technologies are also
increased with diversity, so RECLAIM should also produce greater innovation than the

61
Title IV market. If cap-and-trade should work anywhere, it should work in the
RECLAIM program. RECLAIM is therefore a critical case study in the way in which the
theory of cap-and-trade meets practice (Eckstein 1975; King, Keohane, and Verba 1994).
This dissertation assesses the extent to which RECLAIM produced the predicted
outcomes. However, the focus of this dissertation is effectiveness. A lengthy discussion
of effectiveness follows. To a lesser extent it focuses on efficiency and innovation, but
does so with the understanding that these are drivers of effectiveness and not only
program goals. The most critical contribution of this study is conceptual and
methodological. How does one conceptualize “effectiveness” in a cap-and-trade program
and how does one measure that? The following section describes general conceptual and
methodological issues surrounding cap-and-trade. It then goes on to describe how this
study measure effectiveness as “relative impact”. By measuring the relative impact of the
RECLAIM program vs. the command-and-control alternative it replaced we can more
fully grasp the potential of cap-and-trade programs as an effective tool in reducing
harmful emissions leading to a safer environment and increased human health.

Methodology

All research designs must meet basic criteria of validity and reliability. Yin
(1994) elaborates upon the standard validity and reliability concerns of researchers,
dividing them into four separate concerns. The first, construct validity is the degree to
which key concepts are correctly operationalized. This study relies on widely agreed

62
upon criteria about environmental policy effectiveness, using publicly available data for
the dependent variables. Less clear are the causal factors driving those dependent
variables. A second test is that of internal validity. Internal validity is the feasibility of
the causal relationships that are identified in the analysis. Using published documents
and prior internal and external program reviews, a causal narrative will be able to provide
greater internal validity than a purely quantitative research design in analyzing several
independent variables effects on the dependent variables. Next, external validity is the
capacity to generalize on the basis of a study’s findings. A case study that thoroughly
explains a unique phenomenon but cannot be extended to similar cases is not very
informative. For small N studies like this one, the depth of the analysis can be very
useful, in seeing whether the program, which looks on its face as a best case example of
cap-and-trade, really meets the underlying attributes on which success relies. While its
predictive power may not be especially powerful, it serves as a useful starting point for
further investigation. Finally, reliability is the certainty that the measurement and
findings can be replicated by other researchers (Yin 1994). In the tradeoff between
validity and reliability, the comparative method is generally weaker on reliability because
the small number of observations could reflect the biases of the analyst (Shaffie 2002).
This study makes no claim that the cases studied are a representative sample. The number
of cap-and-trade programs in the U.S. is relatively small making a large N study where
the programs represent a sample of the universe not feasible. The RECLAIM program
has been chosen for the specific reason that it meets the two basic requirements for
predicted policy success.

63
Yin (1994, 15) notes that case studies have a distinctive place in policy research.
Accordingly, he argues that case studies have at least five different uses in evaluating
public policies. First, and most important, is the use of the case study to explain the
causal links in real-life policy interventions that are too complex for other methodological
strategies. Second, case studies may describe an intervention and the real-life contexts in
which it occurred. Third, they can illustrate certain topics within an evaluation in a
descriptive mode. Fourth, they can be used to explore those situations in which the
intervention being evaluated has no clear, single set of outcomes. Last, the case study
can be used as a sort of “meta-evaluation”—what he calls a “study of a study.”
When looking at cap-and-trade markets as a regulatory tool, some methodological
problems are glaringly obvious. The most important problem is the small N involved.
That is, relatively few cap-and-trade markets have gone from design to implementation.
If the medium of the environmental problem to be studied effects the policy tool in use,
then the number of truly comparable policies shrinks even more. While the number of
these markets continues to increase, absolute numbers remain relatively low. Fewer still
have been in existence long enough to really fairly evaluate them. A policy must be given
time for implementation and adjustment before meaningful observations can be made.
For instance, in the first few years of the RECLAIM program several studies proclaimed
it a success on a number of fronts (Kamieniecki, Shaffie, and Silvers 1999; Colby 2000).
Later studies proclaimed the program a failure (EPA 2002; Wright 2006; Zhan 2006). As
we shall see, many of these proclamations were premature.

64
While no comprehensive inventory of cap-and-trade markets is known to exist,
the numbers of large scale programs targeting some form of air-pollution in the U.S. is
less than ten.
1
The largest and most notable of these programs is the SO2 market
designed to regulate Acid Rain precursors as part of Title IV of the 1990 Clean Air
Amendments. RECLAIM, although much more limited in geographic scope, is the only
other large-scale cap-and-trade market in continuous existence for longer than ten years.
Further, on its face it is larger and more heterogeneous than the Title IV program.
To date the Title IV acid rain program has received multiple evaluations,
however, the most thorough has been the Ellerman et. al. study (2000). They chose the
SO2 market as an example of cap-and-trade regulation because it was the largest and
most well developed market yet devised. Further, the market had been implemented long
enough to begin to draw policy conclusions from it. The study had two important
features. First, it examined multiple facets of efficiency and effectiveness of the
program. Second, the study was groundbreaking in its sophisticated uses of quantitative
counterfactuals. Counterfactual analyses ask what would have happened had some
independent historical variable been different. What would the world look like today had
some event been changed? Counterfactuals range from the unsophisticated, such as

1
This is based on my reading of a number of different sources included OECD’s (2002) review of domestic
tradeable permits, Colby (2000), Bryner (1999), and others. The most comprehensive attempt at
inventorying cap-and-trade and other market instruments to date was done by Stavins (2000) for Resources
for the Future. The lead reduction program of 1986 probably qualifies as both large scale and having been
fully implemented for a sufficient length of time, but since lead in gasoline is now illegal, the program has
ceased. State programs at reducing VOCs might meet these criteria and are still in operation, but they have
had predictably limited success. The nature of VOCs is such that they are difficult to monitor and
therefore are not the best candidates for this type of regulation. There is also disagreement in the literature
as to what constitutes a “cap-and-trade” program vs. what is a “credit program” (See Stavins 2000).

65
discussing with a friend how the world would be different had Kennedy not been
assassinated, to sophisticated quantitative path dependency models. By using a number
of sophisticated counterfactual baselines, Ellerman and colleagues asked the important
“compared to what question”. That is, the SO2 market regulation actually implemented
was compared to a hypothetical command-and-control market that could have been
implemented. By using counterfactual analysis, the authors also effectively turned a
single case study into a comparative case study. They compared the actual empirical
results of implementation of the SO2 market with what would have happened had power
producers been forced to adopt smokestack scrubbing technologies.
The case study method is also employed here. RECLAIM was chosen for the
same reasons that Ellerman and colleagues chose the SO2 market. RECLAIM is the
most fully developed regional cap-and-trade market with over 400 firms participating in
the program in any year. This is important because RECLAIM, like the SO2 program,
comes closer to approaching the ideal market than smaller cap-and-trade programs. As
such, using RECLAIM avoids the potential criticism that its deficiencies are not
generalizable because it represents an inappropriate use of cap-and-trade as a regulatory
tool. Further the program is now past the ten year mark giving ample time for
adjustments to unexpected circumstances. Sabatier and Jenkins-Smith (1993) claim that
no policy ought to be studied until it is at least ten to fifteen years old to allow feedback
and adjustments. While this figure might be high, especially given the need for
information in the political process, their point that policies ought to be given some time
before we judge the extent to which they are working is well taken. RECLAIM is both

66
old enough and has gone through enough adjustment to meet this concern. Unlike the
SO2 program, though, the RECLAIM market has gone through considerable price shocks
and considerable rules adjustments. RECLAIM, therefore, has the added feature of
having undergone adjustments to unanticipated events, making the descriptive aspect of a
case study all the more important.
RECLAIM has also received widespread attention from domestic and
international policymakers making it consequential in the real sense envisioned by King,
Keohane, and Verba (1994). For instance a recent OECD report reflects the growing
interest in ex post evaluations of market instruments in implementation:
Evaluation evidence on the performance of emissions trading and other tradable
permit mechanisms in the US could contribute significantly to the international
diffusion of tradable permits, and to more efficient policy learning about the
contexts to which they may be most effectively applied, and the best form for
tradable permit systems to take. (OECD 2002, 31)

As market instruments begin to be diffused on a global scale, policymakers are looking to
RECLAIM as an example of why cap-and-trade strategies ought to be adopted. Further,
RECLAIM was developed against the background of a preexisting Air Quality
Management Plan (AQMP) which envisioned regional air standard compliance through
the use of command-and-control. This AQMP was developed to bring the region in line
with the federal Clean Air Act and with the California Clean Air Act which required the
SCAQMD to develop a plan to reduce air pollution in the Los Angeles basin by 50% –
5% a year for ten years. Assuming that the models used to construct the AQMP plan
were correct, it can be used to represent the answer to the ever important “compared to

67
what” question. While not as rigorous as the Ellerman et al (2000) counterfactuals, the
AQMP can serve as a general proxy for what would have occurred under command-and-
control. This counterfactual analysis is exactly how the SCAQMD measures program
success. When looking at RECLAIM, the implementing agency believes it can see what
its regulatory alternative actually would have looked like.
The Title IV acid rain trading program has been touted as the ideal by which all
other cap-and-trade programs ought to be measured. This serves an important baseline
from which lessons can be learned for future markets, but it is not the only baseline.
When comparing what seems to be a single policy which has not lived up to expectations
with a single ideal, the researcher runs the risk of drawing unreliable inferences. For
instance, there is a risk that the researcher may conclude that since ideal conditions under
which the program will work will never be encountered that the program should not be
adopted. Of course, this overlooks the fact that rarely does a policy work as well as the
ideal and that next best alternative probably will not work ideally either. It is important
to compare cases of successful cap-and-trade markets to failed ones, but it is equally
important to compare each to the command-and-control alternative. As Glazer and
Rothenberg (2001) rightfully note in another context, it is not enough to be critical of the
shortcoming of markets without being equally cognizant of the shortcomings of
government interventions.
If we adopt the view that implementation strategies are highly contingent on
context, then case studies ought to prove useful in finding out just which contextual
variables matter and which do not (deLeon 2001). An in-depth analysis of RECLAIM

68
compared to the command-and-control alternative – as operationalized in a quantitative
counterfactual or in a comparison between two groups –should help explain the causal
links of cap-and-trade implementation that are too complex for other methodological
strategies. Also, the RECLAIM case should also describe this intervention in the real-life
contexts in which it occurred as opposed to those predicted (Yin 1994). Last, RECLAIM
is presently being used as an example of a successfully implemented cap-and-trade
market worthy of emulation on a global scale (OECD 2002). For these reasons, an ex
post evaluation of RECLAIM should prove fruitful in examining real world
contingencies for which ex ante predictions could not have foreseen. On the one hand, it
may be that certain contextual factors were largely overlooked by those who predict a
great deal of policy success from cap-and-trade programs. On the other hand, cap-and-
trade programs such as RECLAIM might prove to be problematic, but the problems
associated with them may be no more than those encountered by alternative regulatory
arrangements. The case study method has the advantage of a much deeper analysis of the
causal mechanisms underlying each regulatory form than a large N study. The in depth
nature of the case study also lets us examine the attributes of the program, as
implemented, in a way that a large N quantitative study might overlook.

Evaluative Criteria

This research seeks to explore the extent to which cap-and-trade markets have
produced certain benefits that might not be found in command-and-control. Because a

69
program can be judged a dismal failure from one perspective but “may well be
considered a ringing success from another” (Kraft 1998, 303), this study adopts a
multiple evaluative criteria framework. From the discussion in the previous chapter, it is
argued that the “success” of any environmental policy rests on at least two factors:
effectiveness and efficiency. These two factors are the core goals of modern
environmental programs. However, as previously argued, these goals should not be
treated as competing or conflicting values in environmental any given environmental
policy. These goals are treated here as hierarchically nested policy preferences in
legislative intent. Effectiveness is a higher order priority in environmental policies than
is efficiency. Cap-and-trade markets were designed to meet the same environmental
targets as the command-and-control regime that they replaced, however they are intended
to do so in a more efficient manner. Since implementation studies seek to connect policy
intent to policy outcomes, effectiveness and efficiency are treated as dependent variables.
The key independent variable looked at in this study is the regulatory treatment.
This is a dichotomous variable: either facilities are in the RECLAIM universe or they are
not. The literature specifies that cap-and-trade should produce effectiveness outcome no
worse than command-and-control, and that the possibility exists that they will do much
better. There is also near unanimous agreement that cap-and-trade should produce more
efficient results.
For the reasons stated in the previous chapter, though, efficiency is also a key
independent variable in producing effective outcomes. An inefficient cap-and-trade
market should not produce effective outcomes comparable to command-and-control.

70
Gerring (2004) offers an additional insight into the usefulness of a case study.
Case studies are not as generally useful in disconfirming hypothesis as are statistical
methods. This is because a single case may not be a representative sample of the larger
universe. Cases are better at producing hypothesis than they are at falsifying them.
However, he notes that there is an exception to the general rule: case studies may
disconfirm a hypothesis when it relies upon a necessary or sufficient condition. One
hypothesis deduced from the literature on the causal mechanism underlying how cap-and-
trade markets drive down emissions is that in order for the mechanism to work properly
there must be an efficient market. An inefficient market is a necessary condition for cap-
and-trade markets to reduce emissions. Is this the case in RECLAIM?

Equity and democratic norms

Other measures could easily be constructed to test criteria external to those
imposed by the legislative language in Title IV and the rules governing RECLAIM.
Perhaps the two most important criticisms of cap-and-trade markets is the argument that
they may be less equitable and less democratic than command-and-control. As argued
previously, to interject normative evaluative criteria other than those specified in
legislative intent is to inject the researchers own bias into the research design.
Implementation studies seek to explain the link between programmatic goals and
programmatic achievements (Knaap and Kim 1998), not to seek to explain the link
between a researcher’s goals and programmatic achievement. Such externally imposed

71
criteria may be important to advocacy groups whose policy preferences are not expressed
in programmatic goals (Schneider and Ingram 1997), but do little to advance the study of
implementation. Therefore, these variables are not considered extensively in this study,
but they do merit some discussion.
For the most part equity criticisms have come in the form of concerns over “hot
spots” of pollution being created (Maclean 1995; Foreman 1998). That is, cap-and-trade
programs may impact certain portions of the population negatively even as total
geographic pollution levels fall. Those that are most likely to be negatively impacted, it
is argued, are usually the poor and minorities. Substantial anecdotal evidence suggests
that there may be some truth to the claim that, in general, minorities receive fewer
benefits from environmental policies. However, one empirical study of RECLAIM finds
that while hot-spots do occur, they do not impact poor and minority areas in any greater
proportion than they do wealthier and predominately Anglo communities (Hawkins et. al
2001).
A second criticism is made by those that concede cap-and-trade markets may be
more efficient than their regulatory alternatives, but that they are highly undemocratic
and therefore undesirable. A major democratic concern is whether market-based
approaches seriously restrict public input in the policy process (Kamieniecki, Shafie, and
Silvers 1999). Since implementation decisions are delegated to private firms, normal
routes of public involvement are circumvented. Another criticism arises out of the fact
that great information obstacles must be overcome for the public to participate because of
the highly technical nature of the regulatory program. A last criticism is that although the

72
public may participate in the system by purchasing allowances on the open market, that
few are able to do so. What is not clear in any of these criticisms is the extent to which
cap-and-trade markets are more undemocratic than command-and-control. In both cases,
public input into the implementation process is done at the rule-making stage, but after
that the public has little say into the day-to-day routines which govern the lives of policy
implementers.
There is also some merit to the argument that cap-and-trade programs might be
more democratic than the command-and-control alternative. Interest groups outside the
universe of targeted facilities are generally able to participate in the market under cap-
and-trade. Any interested individual or group may buy tradable permits, which in the
case of RECLAIM is the “RTC” [Reclaim Tradable Credit]. Since each RTC bought on
the open market by individuals or environmental groups who will not in turn use the
credit to pollute, each of these transactions represents a net gain for environmental
quality. Allowing such participation by the public in the decision over how much overall
yearly pollution will be allowed is unprecedented in command-and-control. And while it
is true that there are many obstacles to be overcome in order for groups to form that are
sufficiently prepared to participate in the market, those same obstacles to group formation
exist regardless of whether that participation is conceived of in terms of the market or the
rulemaking stage (Moe 1991).
While these concerns raise interesting questions about the ability of cap-and-trade
markets to meet normative goals paramount to democratic governance, they are excluded
in this analysis as major dependent variables. Neither meets the criteria of being

73
programmatic goals and therefore falls outside the scope of this study. In the narrative in
the following chapters, both concerns will be addressed, but only tangentially. Each
deserves a much deeper analysis than will be offered here.

Operationalizing effectiveness

The following describes the main dependent variable explored in this study:
policy effectiveness. Gormley (2002) categorizes and assesses four types of effectiveness
measures:
Outputs are measures of what government agencies do on a day to day basis, such
as inspections, the issuance of permits, the provision of technical assistance, or the
imposition of monetary penalties. Using outputs as an indicator of effectiveness has
received widespread criticism in the environmental policy literature (Bartlett 1994).
Outputs are probably used so often because these data are easily obtainable, especially in
agency sponsored evaluations (Weiss 1972). Though measurable, outputs do not tell
policymakers or managers whether progress has been made towards achieving stated
goals. At its worst, output studies amount to “bean-counting” with bureaucratic routines
displacing statutory goals as values to be maximized (Gormley 2002, 14-15; Kraft 1998).
Intermediate outcomes are measures of the environmentally relevant behavior of
regulated firms or citizens. Intermediate outcomes are often used because the
relationship between the activities of targets of regulation and environmental quality is
well documented. As outlined in chapter one of this dissertation there is a great deal of

74
evidence suggesting that the emissions of NOx and SOx have some adverse effect on
environmental quality and, in turn, on human health. SOx is a precursor to acid rain and
NOx is the major contributing factor to ground level ozone, otherwise known as “smog”.
A decline in those emissions represents a valid measure of environmental improvement.
The possibility remains, though, that intermediate outcome measurements can be
just as poor an indicator as outputs if based on an invalid theory of the factors affecting
the policy problem (Mazmanian and Sabatier 1983). For instance, compliance rates are
an intermediate outcome measure. Compliance rates are only a good proxy for
environmental effectiveness if the underlying regulations which are being complied with
can be shown to have a causal relationship with eventual environmental quality. If the
causal relationship is weak, compliance rates are no better than output studies.
Intermediate outcomes are superior to output measures when they focus on the activities
of targeted groups which adversely affect the environment.
Terminal outcomes are measures of trends in environmental conditions or human
health. Terminal outcomes are the overall goals of environmental policies – cleaner air,
fishable or swimmable water, less waste in landfills, better human health, etc. While
terminal outcomes are perhaps the ultimate reason for enacting a policy, nevertheless
they are not often used in measuring policy effectiveness. Terminal outcomes can
sometimes be difficult to measure. Trends in human health, for instance, can be
extremely difficult to ascertain. Some outcomes may be expensive or time consuming to
measure. Watershed quality studies, for instance, are data intensive and difficult to do.

75
Further, even when data are measurable and available, they may not be the appropriate
proxy for policy effectiveness.
Terminal outcomes are more useful when assessing a bundle of environmental
policies rather than any single policy instrument or implementation strategy. This is
because any changes in environmental quality can only partially be attributed to any one
program. For instance, in Los Angeles air quality has improved over the last 20 years
(SCQAMD 2002), but it would be incorrect to claim that any one program was
responsible for the change. A number of different policies, programs, and socioeconomic
trends all contributed to pollution reduction in the region. RECLAIM is only one
program among dozens of other programs, rules, and laws that have contributed to the
empirically valid observation that the air is cleaner. Thus, while terminal outcomes are
important measures of overall environmental quality, they are not usually appropriate
when analyzing a single policy which is only part of a larger bundle of policies affecting
the change.
Impacts are measures of the effects of outputs on outcomes. Gormley (2002, 15)
defines them as “the marginal contribution of administrative activity to something of
value, such as lower industrial emissions, or improved air quality.” Impacts can be
measured as change in intermediate or terminal outcomes. Impacts are the holy grail of
policy studies—alluring and elusive. They are alluring in that they measure exactly what
most policy studies want to know: how much is this policy is contributing to that
outcome (Knaap and Kim 1998)? Impact measurements have been used successfully in
such areas as education and health care (Gormley and Weimer 1999). Impact

76
measurements, though, remain elusive in environmental policy studies because it is
difficult to disentangle the effects of administrative activity, economic activity, highly
erratic meteorological phenomena, entrepreneurial innovation, cultural change, and other
factors on environmental outcomes, whether intermediate or terminal (Gormley 2002).
Nevertheless, impacts are the central question to policy studies in general, and
implementation studies in particular (Meir 1999).
With these four aspects of effectiveness in mind, which is the most appropriate in
thinking about cap-and-trade programs? This study takes the position that impacts should
be the central theme of policy studies. The underlying theory of Title IV, RECLAIM,
and all other cap-and-trade programs is that by reducing the amount of pollution
produced by participating facilities, overall environmental quality will be improved. In
the case of Title IV it is the reduction of acid rain. In the case of RECLAIM it is the
reduction of smog and other air pollution in the Los Angeles basin. However, the
marginal contribution of a single program to overall air quality is beyond the scope of
this research. It is not possible to measure the exact affect each of these programs has on
ultimate goals. Multiple pollution sources not regulated under Title IV exacerbate acid
rain. In the Los Angeles area it is probable that federal, state, and local regulations
governing automobile tailpipe emissions — which are not included in the RECLAIM
regime — have been the biggest contributors to smog reduction in the basin.
Impacts of cap-and-trade programs on certain intermediate outcomes are possible
to measure and data are readily available. Cap-and-trade markets are an effort to change
the polluting behavior of certain targeted populations. Because the underlying theories

77
that these targeted behaviors do affect the ultimate environmental qualities in question,
those changes are considered an appropriate intermediate outcome here.
For RECLAIM, the target behavior is the release of NOx, a major precursor to
smog formation, and SOx, a major contributor to acid rain. The impact of the RECLAIM
program is the change in the outcome of the total amount of NOx and SOx gasses at
participating facilities, all other things being held constant. The net effect of the program
is the impact it has had on changes in overall emissions levels by participating facilities,
all other things being held constant.
This is similar to studies of the Title IV program which have included an impact
variable where the targeted behavior is the release of SO2, a major contributor to acid
rain (Ellerman et al 2000). The impact of Title IV is the change in the outcome of the
total amount of SO2 gasses at participating facilities, all other things being held constant.
Assuming that the underlying causal theory is correct, this relationship can be expressed
by:

Impact = Outcome at T2 – Outcome at T1 (all other things held constant)

The effectiveness of air-pollution regulation can be measured in three ways. First,
they may be effective to the extent that the impact of the program positively reduces the
outcome of total targeted pollutants. The purpose of any environmental program is to
change actual environmental conditions. If a cap-and-trade market program actually

78
lowers the amount of total emissions at participating facilitates, it can be said to be
effective. Thus:

If impact > 0, then effective.
but
If impact ≤ 0, then ineffective.

This measure operationalizes the concept when effectiveness is not wholly dependent on
programmatic goals. For instance, command-and-control regulations often failed to meet
aggressive pollution reduction targets. But would it be fair to say that these programs had
no impact, or that they were ineffective if those goals were simply unrealistic? Such
programs might be labeled less effective than hoped, but it would be inaccurate to label
then ineffective.
Second, air pollution regulation can be said to be effective when the impact of the
program is such that a reduction of targeted pollutants at participating facilities meets
programmatic goals. This is the approach usually taken in the implementation literature,
and one of the measures adopted here. Thus:

If impact ≥ program goal, then effective

The programmatic goals in both the Title IV and RECLAIM program have been
operationalized by the EPA and SCAQMD, respectively, by the total number of pollution
permits allocated in each program. Cap-and-trade markets have an inherent advantage

79
over command-and-control in that set pollution reduction goals are very clear. If the
ultimate goal of the program is to reduce emissions to X amount, then X amount of
permits are issued. Unless permit holders violate the terms of their permits there will be
no excess pollution. Under command-and-control technological fixes are required based
on the theory that these fixes will lead to an overall reduction in emissions. The exact
amount of emitted pollution, however, will depend on variances in the amount of
production at the regulated facility. For instance, a ‘scrubber’ might be installed at a
power-plant to reduce total output of SO2. The exact amount of SO2 released by the
plant will fluctuate depending on the intensity of plant usage. During unseasonably warm
or cold years, a power plant might be used much more intensely than planned. If
emissions at a targeted facility exceed the expectations of regulators, there is no legal
recourse since the technical requirements of the law have been satisfied. Under cap-and-
trade, the ultimate goal of the program is to limit pollution to the number specified in the
tradable permit.
Under this definition of effectiveness, if the total pounds of pollution by
participating facilities are equal to or greater than the total number of permits allocated to
them, the program is effective. For both programs, permits given to participating
facilities are expressed in pounds of pollution that may be released. Thus, for both Title
IV and RECLAIM, the effectiveness relationship can be expressed:

If lbs. of targeted pollutants released ≤ total permits, then effective

80
This is one of the measures of effectiveness employed in this study. Successful program
implementation can be determined by the extent that outcome measures, as specified by
the EPA and SCAQMD, meet or exceed program target goals.
The third and the most useful effectiveness measure is to compare the actual
impacts of a program to those of the next likely regulatory tool. This is relative impact.
Effectiveness is best conceived of as a function of how well a program achieves defined
policy goals compared to other program alternative which also seek to achieve the same
set of defined goals. Cap-and-trade markets may be effective, but are they effective as
the command-and-control regulations that they replaced? Under this definition, cap-and-
trade markets can be said to be effective when the impact of the program is as great as the
impact of its regulatory alternative (Ellerman et al. 2000). Generally, then:

If impact ≥ regulatory alternative, then effective

This dissertation uses two methods for measuring the relative impact of RECLAIM vs.
the command and control alternative: counterfactual and comparative analysis.

Counterfactual construction

Ellerman and colleagues (2002) construct three quantitative counterfactuals which
they use as a baseline from which to compare actual Title IV results. While the literature
on cap-and-trade markets emphasizes that this particular implementation strategy is

81
especially appropriate when regulatory targets are heterogeneous, the Title IV program
targets the relatively homogenous group of coal-fired power plants. In order to meet their
permit allocations, power producers basically had the binary choice of either switching to
low sulfur coal or install “scrubbers.” Thus, their study was able to compare the results
of a cap-and-trade market to what would have been produced under a command-and-
control regime in a relatively straightforward manner. Further, while they do consider
relative impact to be an important dependent variable, their primary concern is with the
relative efficiency of the program.
Markets such as RECLAIM are different in that the targets of regulation are much
more diverse and therefore have a wider range of options to choose from in reducing
emissions. Because RECLAIM meets the heterogeneity criteria better than Title IV
constructing counterfactuals as reliable as those used by Ellerman et al. (2000) are much
more difficult. Estimates of what would have happened under command-and-control,
though, were made by the SCAQMD in the original AQMP (SCAQMD 1997). This may
serve as a rudimentary counterfactual by which to compare RECLAIM performance with
its command-and-control alternative. The AQMP model represents the expectations of
regulatory officials as to impacts of proposed command-and-control regulations.
A major caveat is that the models used to produce the likely pollution path under
the original command-and-control AQMP are proprietary and not publicly available. We
are informed by SCAQMD Rule 2002 which specifies that RTCs will be issues based on
the maximum emissions throughput from each pollution source from 1989 through 1992,
but a private firm was actually contracted to come up with facility by facility permitting

82
levels. This calls into question the underlying assumptions of the counterfactual model.
Further, ex post empirical observation of facilities in the region that continued to operate
under command-and-control during the same time period as RECLAIM suggest that the
counterfactual baseline produced by the SCAQMD is not a valid measure. This study
will return to the theme of the validity of the AQMP model in the next chapter.
However, given the inherent limitations of the AQMP model, it is still useful in
comparing RECLAIM impacts to the predicted impacts of the command-and-control
alternative given that it is the measure of relative impact used by the implementing
agency.
Under this theory the total number of tradable permits given to participating
facilities in the RECLAIM program represents the total amount of pollution that would
have occurred under the AQMP command-and-control alternative (SCAQMD 1997).
Thus, the cap-and-trade market of RECLAIM can be said to be as effective as the
command-and-control alternative when total emissions are equal to or less than the total
amount of emissions allowable under the terms of the program. Thus:

If total lbs. emissions ≤ permitted total lbs. emission, then effective

Under this definition, even when compliance rates are less than 100% -- meaning that
some firms exceeded allowable emissions without purchasing additional emissions
credits on the open market – if the total amount of emissions is lower than the total
permitted, RECLAIM would be considered at least as effective as command-and-control.

83
The number of permits represents a counterfactual baseline by which to make
comparisons to and to draw inferences from. But this is only a valid measure inasmuch
as the predictions made by those formulating the cap-and-trade market represent a
believable alternative universe without the intervention. In other words, does the “cap”
in cap-and-trade really represent the amount of pollution which would have occurred if
the program had not been adopted? If yes, then the number of tradable permits used is a
good proxy for relative impact. If no, then the measure is useless in comparing the
relative impacts of cap-and-trade with command-and-control. Again, we will return to
this in the next chapter.

Comparing CAT intervention to CAC

A second, and what I propose is a more valid, measure of relative impact is
comparing those participating in the program vs. those not participating in the program.
Such a comparative approach more closely resembles that of an experimental design in
drawing inferences from dependent variable variation when one or more independent
variables is introduced into one population but not in the other. In the case of RECLAIM
the treatment introduced is the program itself. Before the RECLAIM program was
instituted, all facilities under the SCAQMD jurisdiction were regulated by various
command-and-control instruments. This is the baseline, or pre-intervention group, which
includes the entire universe of regulated point source of pollution facilities in the greater
Los Angeles basin.

84
In 1994 the RECLAIM program was instituted. All facilities in the SCAQMD
region which emitted more than 4 tons of the targeted pollution were included in the
original RECLAIM universe. Other facilities with lower emissions levels could
voluntarily enter the program. This is the treatment group which received the cap-and-
trade intervention. Those included in the group are not a representative sample of those
participating in the command-and-control program. They are the entire universe of
participants. Thus the validity of whether or not inferences about the RECLAIM
program can be drawn from these data is not in dispute. No claim is made that the case
of the RECLAIM intervention group is necessarily reflective of the larger phenomenon of
cap-and-trade programs. The case was chosen precisely because it seemed like a likely
candidate for a market based program which should produced the desired affects.
The third group is made up of all those facilities which continued to operate after
the 1994 cap-and-trade intervention but which for various reasons were not included in
the RECLAIM universe. This is the control group. It is proposed that this group
represents a better baseline from which comparisons can be made in assessing relative
programmatic impacts. This method is superior to counterfactual analysis when the
counterfactual is based upon ex ante predictions of programmatic impact on an affected
population. Ex post evaluation of how those in the control group reacted to continued
participation in a command-and-control environment is informative on two fronts. First,
it tests the validity of the counterfactual baseline used by the implementing agency.
Second, it informs us of the relative changes which those in the treatment (RECLAIM)
might have experienced had no intervention taken place.

85
One can draw meaningful inferences about the relative impact of RECLAIM by
comparing the impact the RECLAIM program had on emissions levels vs. the impact of
command-and-control on those facilities not participating in the program. One can
compare the relative impact of RECLAIM by comparing variation in year to year
emissions between the RECLAIM and non-RECLAIM control group. If the intervention
significantly altered the behavior of those in the RECLAIM group, then we should see
substantial year to year variation between those participating and those not participating.
But perhaps a more meaningful measure of the relative impact of the RECLAIM
program would be to see how those facilities included in the program changed their
behavior over the lifespan of the program. This requires a comparison of emissions
levels between each of the two post-intervention groups and the baseline pre-intervention
universe. If the RECLAIM program is a successful intervention then we should see
emissions levels dropping from year to year relative to the pre-intervention start of the
program. We should also expect to see that emission levels changes among those
participating in the program vary substantially from those not participating in the
program. If emissions levels in the intervention and the control group tend to change
together, then it makes sense that the causal driver in variation is not the intervention.
What must be assumed is that some other external factor is driving dependent variable
change, and that the independent variable in question [inclusion in the cap-and-trade
program] has no effect.
All of these measures of effectiveness are adopted in this study as all are based on
explicitly or implicitly stated goals in the RECLAIM program. First, RECLAIM, like all

86
air-pollution regulations, was adopted with the explicit goal of lowering emissions of
certain ambient pollutants at participating facilities based on the valid theory that these
emissions harm the environment and have damaging effects on human health. Second,
RECLAIM was established with the explicit goal of lowering those emissions by specific
amounts. Third, RECLAIM was established with the explicit goal of lowering those
emissions by at least the same amount as what would have occurred under command-
and-control. The two methods of measuring this last program goal, defined here as
relative impact, are useful in trying to assess the program’s success.

Efficiency as an independent variable

The goal of cap-and-trade markets is to provide equal or greater reductions in
pollution than the command-and-control alternative in a more efficient manner. The
literature on the creation of marketable goods predicts greater efficiency than command-
and-control because both the buyer and the seller are better off as a result of the
transaction (Weimer and Vining 1989). The two parties would not consent to the
transaction in the first place without expectations of greater marginal utility. Under cap-
and-trade markets polluting firms can buy and sell pollution permits giving them the
flexibility to determine for themselves the best way to meet their emissions limit, which
is determined by the number of pollution credits they own. A firm may choose to pay for
new equipment to reduce its pollution level or it may purchase additional credits from
another firm—whichever is cheaper (Kamieniecki, Shafie, and Silvers 1999, 113). Two

87
other, less discussed, options are available to the firms. Because a single firm may own
multiple targeted plants, a firm may choose to reduce emissions at one plant and transfer
credits, without price, to another plant. Finally, a firm may simply suspend normal
operations when the price of tradable credits is high enough. In this last scenario, a
smaller firm with a large amount of emissions credits might find that it can make larger
profit by selling all of its credits on the open market than it did producing goods.
Under these conditions, the marginal cost of pollution control is equalized across
all firms. It is in the interest of each polluter to cut emissions to the socially efficient
level, where the price of an additional emissions credit is equal to the marginal cost of
abatement. After this, any increase in emissions would reduce the firm’s profit (Hahn
and Noll 1982). Firms that are able to reduce their emissions beyond those permitted
may sell their unused credits on the open market. It is also argued that the cap-and-trade
market method encourages technological innovation by giving firms an additional
financial incentive to find cheaper ways to reduce pollution. Under command-and-control
regulation the technology that must be used for compliance is stipulated, but there is no
incentive to more efficiently reduce pollution levels beyond those prescribed by law.
Trading is key to the efficiency of cap-and-trade markets. The existence of a competitive
market makes it possible for firms to meet specified emissions levels at a lower overall
cost than is possible under direct command-and-control regulation (Kamieniecki, Shafie,
and Silvers 1999, 113).
Efficiency is defined in the public policy literature as the ratio between inputs and
outputs. In environmental policy this means that “the stated environmental quality

88
standards are achieved at the lowest possible costs” (Freeman 2003, 209). When
assessing cap-and-trade policies, academic disciplines often tend to emphasize certain
aspects of efficiency over others. For instance, economists often focus on the efficient
operation of the market for tradable permits (Colby 2000; Johnson and Pekelney 1996;
Joskow et. al. 1998). For public administration, concern tends to be with compliance,
monitoring, and enforcement costs. The best evaluations, though, should take into
account as many costs as possible. The OECD (2002, 36-37) has suggested that at least
three aspects of efficiency that should be particular priorities in ex post environmental
evaluations: market efficiency, administrative efficiency, & compliance efficiency

1) Market efficiency. Market efficiency occurs when the marginal cost of pollution
control is equalized among all firms. Tietenberg’s highly influential Emissions Trading
(1985) gives the most straightforward explanation of why cap-and-trade markets ought to
be efficient:
Plants have very different costs of controlling emissions. When credits are
transferable, those plants that can control most cheaply find it in their interest to
control a higher percentage of their emissions because they can sell the excess.
Buyers for these reductions can be found whenever it is cheaper to buy emissions
reduction credits for use at a particular plant than to install more control equipment.
Whenever an allocation of control responsibility is not cost effective, further
opportunities for trade exist. When all such opportunities have been fully exploited,
the allocation is cost effective (p. 16).

The market is the place where the buyer and seller of pollution permits come together.
Since pollution credits will only be bought and sold when there is economic incentive to
do so, markets can be said to be efficient when the price of a pollution permit reflects all

89
the information necessary to make an informed buying or selling decision. Relevant
indicators of market functioning might include volume of trades, the numbers of buyers
and sellers, and the extent of price dispersion between trades (OECD 2002, 36). To the
extent that variance in these indicators is minimized, a market can be said to be more
efficient (Tietenberg 1985).
According to Hahn and Noll (1982), at least three conditions are necessary for
markets to work efficiently. Kamieniecki, Shafie, and Silvers (1999, 113-114) adopt
these conditions in their study of RECLAIM and they are adopted here. Two other
conditions are adopted and based on the work of Glazer and Rothenberg (2001). The
extent to which RECLAIM meets these criteria of necessary for a market to be efficient
in light of the implementation literature is explored in more detail in chapter five. The
data presented in that chapter helps illuminate the process by which cap-and-trade in
theory becomes cap-and-trade in practice and how many of the assumptions made about
RECLAIM have overlooked important program details.
1a) Adequate market participation (Kamieniecki, Shafie, and Silvers 1999,
113). For a market to be fully functional, enough buyers and sellers must participate so
to avoid the problems associated with monopolistic or oligarchic behavior. Because
prices in the market reflect marginal abatement costs participation is key to sharing firm
level cost-effectiveness information. If too few participants are involved in too few
trades, this reduces the probability that credits will be traded at market-clearing prices.
The number of market participants for RECLAIM is publicly available. As
previously discussed, on its face RECLAIM has more market participants than Phase I of

90
the Title IV program. What is interesting about the literature is that while there is general
agreement that there must be an adequate number of participants in the market, no study
has yet to operationalize this concept. For instance, Kamieniecki, Shafie, and Silvers
(1999, 13), after looking at the number of participants and the number of trades in the
RECLAIM program come to the conclusion that the condition of an adequate number of
participants has been met. Ellerman et al (2000) come to the same conclusion with a
smaller number of participants. There may very well be a sufficient number of
participants in the market, but the literature is not clear as to below what number
insufficiency occurs.
Market efficiency in cap-and-trade markets is different then other measures
heretofore addressed because it is both a dependent variable and a causal driver. If cap-
and-trade markets are to be effective and efficient then a minimum requirement is market
efficiency (Tietenberg 1985). Market efficiency is an explicitly stated goal of the
RECLAIM program and thus a dependent variable, for the purposes of this study we will
focus on its causal relationship with the variable in question: effectiveness.
1b) Low transaction costs (Kamieniecki, Shafie, and Silvers 1999, 114). The
most direct form of transaction costs in cap-and-trade markets come through regulatory
oversight. Some pollution trading markets have encountered high transaction costs
because they require lengthy regulatory review before a market transaction can proceed
(Ellerman, et. al 2000). Another key to keeping transaction costs low is the emergence of
transparent trading mechanism where buyers and sellers can easily attain price
information. Imagine a capital market where buyers and sellers of stock had to find each

91
other without a place of exchange, such as the NYSE. Some cap-and-trade markets have
encountered shaky starts because of lack of foresight into exactly how buyers and sellers
would be brought together. The emergence of either public or private brokerages is an
important step in lowering the cost of participation. Further, government can reduce the
transaction costs associated with information gathering through full public disclosure of
trades, prices, and compliance (Colby 2000). The OECD (2002) suggests that ex post
evaluations can be particularly useful in showing how transaction costs can inhibit rosy
ex ante predictions.
1c) Sufficient monitoring and enforcement (Kamieniecki, Shafie, and Silvers
1999, 114). A cap-and-trade market quickly becomes meaningless if users can exceed use
limits with impunity. The purchase of a pollution credit is the acquiring of a property
right and it is government that must enforce these rights. Lax monitoring or weak
enforcement is a disincentive for participants to acquire new credits because the value of
these credits cannot be assured over time. On the one hand, monitoring and enforcement
must be sufficient to ensure faith in the long term integrity of the rights being purchased.
On the other hand, monitoring and enforcement costs cannot be so excessive that they
defeat efficiency goals altogether.
1d) Credibility. As illustrated in chapter two, agencies charged with program
administration must not only be committed to enforcing rules (condition 1c above), but in
maintaining them. Glazer and Rothenberg (2001) call this “credibility”. Implementing
agents must find a suitable balance between adjusting program rules when those rules are
found to hinder program goals, and maintaining credibility that the program will continue

92
along a predictable path enabling participants to plan for the future. This is especially
important in programs such as cap-and-trade where targeted firms are expected to make
present investments in abatement technologies based on the expectations that the rules
will not be changed in such a way as to alter investment decisions.
1e) Commitment to property rights. As illustrated in chapter two, a necessary
condition for an efficient and therefore effective cap-and-trade market is regulatory
commitment to property rights. In a cap-and-trade market the tradable permit is the right
to emit a given amount of pollution. A tradable permit, in theory, meets all of Dale’s
(1967) conditions of property. In order for a cap-and-trade market to work efficiently,
there must be faith that the property right vested in the tradable permit is secure. Without
this market participants will be reluctant to buy and sell since they have little reason to
believe that the value of the tradable permit will be maintained. Is the SCAQMD
committed to property rights? Chapter five explores this question more fully.

2) Administrative efficiency. Administrative costs are incurred by regulatory agencies for
emissions measurement, monitoring, and regulating permit transactions (OECD 1997).
They are the overall direct cost to the taxpayer for the program. Administrative costs can
sometimes be difficult to measure if agency staff are assigned to several programs thus
making it difficult to disentangle the costs of administering one program from another.
When bureaucratic accounting practices require program costs to be calculated it makes
the job of independent program evaluation easier. Ex post evaluation of enforcement is
especially important in that little can be known about enforcement costs except through

93
observation of actual experience (OECD 2002). However, because internal auditing data
are not publicly available, only a cursory discussion of administrative efficiency is
attempted through a content analysis of publicly available documents.

3) Compliance efficiency. “The cost of compliance is the additional cost incurred to
comply with some environmental or other legal requirement” (Ellerman et. al. 2000,
223). Compliance costs include administrative and managerial costs to the firm in
meeting regulatory standards. When regulatory standards are process outputs,
compliance costs may be extremely high and produce few beneficial terminal outcomes.
In that case “complying” means meeting the technical directives of bureaucratic agencies
and any money spent represents an unproductive use of resources (OECD 2002).
Compliance costs also include the costs to firms for the installation of equipment
designed to reduce pollution. Under command-and-control calculating the cost of plant
equipment is a straightforward task because the cost of equipment remains constant even
across a set of heterogeneous firms.
Under cap-and-trade regulation, a firm is not required to meet any process output
standards. Indeed, the entire justification for choosing market forms of regulation over
other alternatives is that firms should be given discretion in the ways in which they will
meet outcome standards. Further, under cap-and-trade regulation no plant equipment is
required to be installed at the firm level. This makes it somewhat difficult to calculate
the costs of compliance directly under cap-and-trade. Ellerman and colleagues study of
the national SO2 market did attempt to estimate compliance costs, but that program was

94
designed to regulate a single industry—power producers. As already discussed, power
producers had the binary choice of either switching to low sulfur coal or install
“scrubbers”, thus, compliance cost estimations could be made. Markets such as
RECLAIM are different in that the targets of regulation are much more diverse and
therefore have a wider range of options to choose from in reducing emissions. This study
looks at compliance efficiency in light of a content analysis of publicly available
documents.

Conclusion

Efficiency is a goal (dependent variable) of cap-and-trade markets. It is also a reason
(independent variable) why these markets should produce relative impacts at least as
much as their command-and-control alternative. A more efficient market should also
provide incentives for firms to produce greater relative impacts than command and
control and lead to more innovation. If following conditions are met, as outlined in the
above discussion, a cap-and-trade program like RECLAIM should produce effective
outcomes as defined here as relative impact. They should also drive innovation in
abatement technologies:

1) An efficient market, which must include
a) Adequate market participation
b) Low transaction costs

95
c) Sufficient monitoring and enforcement
d) Credibility
e) Commitment to property rights
2) A high level of administrative efficiency.
3) A high level of compliance efficiency.

The greater RECLAIM meets the conditions listed above, the more likely it will produce
relative impacts on emissions levels greater than those found in command-and-control. A
thorough and rigorous quantitative analysis of RECLAIM efficiency is beyond the scope
of this dissertation. An in depth qualitative analysis, using quantitative data where
appropriate, of RECLAIM efficiency is offered in chapter five. A closer look at
RECLAIM shows that, contrary to previous findings, it probably less efficient than had
been previously assumed. An implementation framework is useful in understanding why
RECLAIM does not meet some of the basic conditions for cap-and-trade market success.
The data on efficiency and innovation presented in chapter five help explain the
data and analysis offered in the following chapter. Chapter four constitutes the backbone
of this dissertation and looks at the relative impact of the RECLAIM program. The most
important finding is that the way in which you measure relative emissions levels directly
effects the inferences drawn from those measurements. Using one measure of relative
impact, the one used by the regulating agency itself, one finds that in the vast majority of
years looked at in this study the RECLAIM program has greatly outperformed the
command-and-control regulations it replaced. This seems to confirm the predictions

96
made by the literature since the RECLAIM program better fits the criteria for success in
that it is larger and more heterogeneous than the Title IV program, often touted as a
model of cap-and-trade success. However, a closer look at the counterfactual baseline
use by the SCAQMD shows that it is not a valid measure of command-and-control
performance. Hence, the inference about cap-and-trade is invalid. Using comparative
measurement, one finds that the RECLAIM program is in fact no better at reducing
emissions than those in the command-and-control group. This suggests that either cap-
and-trade programs are not as effective as once thought, or that the RECLAIM program,
as implemented, is not the model of cap-and-trade as many previous studies suggest.

97
Chapter 4 Case Study

Cap-and-trade programs hold the promise of three benefits. First, they may lower
emissions to the same extent as command-and-control; with the additional benefit that
they may lower pollution at an even greater rate. Second, economic theory predicts that
they will be more efficient. Third, they spur innovation in pollution abatement
technologies and processes. Chapter four evaluates the first of these claims: the extent to
which RECLAIM is as effective as its regulatory alternative. As described in the
previous chapter, effectiveness is defined here as the relative impact of the program
compared to the next most likely regulatory alternative. How one looks at effectiveness is
very much dependent on assessment measures. Measuring the relative impact of
RECLAIM is far more complicated than previous studies have assumed. The key
contribution of this dissertation is presented in this chapter. It shows that baselines for
measurement are critical in evaluating public policy performance. If measurement
baselines are invalid, then the inferences drawn from comparisons based on them are also
likely invalid.
This section describes the AQMP that RECLAIM replaced and which has been
used by the SCAQMD as a counterfactual baseline. Using the SCAQMD counterfactual
as a case with which to compare the RECLAIM program, we can see the extent to which
the program has impacted SOx and NOx emissions. Even if the counterfactual is flawed,
it is the baseline by which the implementing agency uses as a measure of success. This
section then explores criticisms of the SCAQMD counterfactual as a valid baseline. A

98
comparative case study between those facilities in the RECLAIM universe and those
facilities still regulated by command-and-control is proposed as an alternative to an
unrealistic counterfactual. Data showing changes over time in emissions levels for both
the RECLAIM and non-RECLAIM universe are presented. These data show the extent
to which RECLAIM is as effective as the command-and-control alternative.

The SCAQMD counterfactual

Annual SCAQMD RECLAIM audits from between 1995 and 2006 were collected
and reviewed. Each yearly audit is prepared by SCAQMD staff and uses data collected
from individual RECLAIM facilities on the number of Reclaim Trading Credits (RTCs)
allocated, traded, and used. Each audit provides a yearly snapshot into the behavior of
RECLAIM facilities. Audits and other RECLAIM documents are publicly available
through the SCAQMD website (http://www.aqmd.gov/reclaim/index.htm). Most audits,
though, cover a time period two years prior to the writing of the audit. So the audit done
in 1997 is actually an audit of the 1995 calendar year and so forth. Every three years a
more thorough program review is required. These program reviews replace the annual
audit but are virtually identical to them. From time to time yearly audits also include
more recent data. For instance, the audit done in 2001 included data from the prior year.
This reflects growing concern in that year over rising RTC prices and the need by policy
makers to have an immediate analysis of the RECLAIM market.

99
The designers of RECLAIM explicitly state that the program should be compared
to the command-and-control system which would have been implemented had
RECLAIM not taken its place. Language reflecting the underlying assumption that a
command-and-control counterfactual exists can be found peppered throughout annual
audits of the RECLAIM program prepared by SCAQMD staff. For instance, the staff
evaluation of the program through 1995 says that (emphasis mine):
RECLAIM sources be required to reduce their emissions to a level equivalent to
the reductions that would have occurred under the subsumed rule and control
measures. Therefore, RECLAIM continues to conform with the requirement for
emission reductions equivalent to the subsumed rule and control measures as
required by Health and Safety Code Section 39616. (SCAQMD 1997 chapter 2, p.
6)

A decade later similar language is used showing a continued faith by the SCAQMD in
the original command-and-control counterfactual (emphasis mine):
The annual allocations given to RECLAIM facilities reflect the required emission
reductions mirroring the reductions anticipated under command-and-control
rules. (SCAQMD 2007 chapter 2, p. 6).

The original number of RTCs allocated to RECLAIM facilities reflects the total amount
of pollution SCAQMD predicted that each facility would be allowed to produce under the
command-and-control alternative. Thus, the total amount of RTCs is equal to the amount
of pollution that was predicted would have occurred.
RTC allocation is synonymous with pollution levels in the counterfactual and
those levels are used by the SCAQMD as a baseline of comparison. In years where the
total number of RTCs actually used was less than the total number of RTCs issued, staff
claimed that this reflects program impact success at least equivalent to what would be

100
expected under command-and-control. For instance, this is how the 1996 Annual
Compliance Year Report comes to the conclusion that pollution under the RECLAIM
program was comparable to what would have been under command-and-control
(emphasis mine):
Allocations are based on the emission reductions which were projected to be
achieved through implementation of the traditional rules and control measures
identified in the AQMP but subsumed by RECLAIM…. Emissions in the 1996
compliance year were well below the total RTC supplies… RECLAIM facilities
did not exceed their allocations on an aggregate basis during the third compliance
year, successfully achieving program emission reduction goals and demonstrating
equivalency to traditional command-and-control measures (SCAQMD 1998, 3 -
2).

The inference drawn from these statements, and many more like them, is that the
effectiveness of the pollution reduction goals of the RECLAIM program should be
judged by the extent to which RTCs are used. RTC usage below the total allowed is
viewed as reduction of pollution over and beyond what would have happened under the
command-and-control counterfactual.
I will call this the “SCAQMD counterfactual” which is what the SCAQMD uses
as a baseline for comparison between the effectiveness of the RECLAIM program and
what would have occurred under command-and-control.

Assessing RECLAIM impact vs. the SCAQMD counterfactual

Throughout the history of the program’s internal evaluations the SCAQMD
counterfactual is widely used as a baseline by which to compare RECLAIM’s

101
effectiveness. Let us first assume that the SCAQMD counterfactual is an accurate
portrayal of what emission levels for the effected facilities would have been had
traditional command-and-control regulations remained in place. As we shall see later,
this is a big assumption and is probably wrong. However, it is the one made by the
SCAQMD and therefore is a good starting point for evaluating RECLAIM’s impact. It is
also the metric by which the agency charged with implementing this market based
environmental policy uses to measure success. It is therefore important if for no other
reason than to see if the program is meeting the internal expectations and goals of the
implementing agency. Using internal evaluative criteria as a measure of how a program
is doing has long been suggested as a way to minimize researcher bias. While using
internal expectations of program performance may not be sufficient in measuring
programmatic success – especially of those expectations are based on faulty premises – it
is a good starting place. This measure is our first test the relative impact of RECLAIM.
Using the SCAQMD counterfactual as a baseline for comparison, how effective is
RECLAIM at reducing the targeted pollutants?
Figure 4.1 and 4.2 show the total amount of reported emissions for SOx and NOx
from 1989 to 2003. The RECLAIM program was not adopted until late 1993 so the first
full year it was in operation was1994, but after that the line showing reported emissions
can be used to visually compare actual emissions vs. what the SCAQMD believed would
have occurred under command-and-control.

102

Figure 4.1 RECLAIM SOx emissions vs. RTC supply 1989 - 2003

Figure 4.2 RECLAIM NOx emissions vs. RTC supply 1989 – 2003

Source: SCAQMD 2007

0
2,000
4,000
6,000
8,000
10,000
12,000
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Compliance Year
SOx (tons)
Total RTC Supply
Reported Emissions
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Compliance Year
NOx (tons)
Total RTC Supply
Reported Emissions

103
Using the SCAQMD’s own standard for judging RECLAIM effectiveness, the
first six years of the program’s life was a resounding success. Between 1993 and 1999
annual emissions of both NOx and SOx were below or equal to the SCAQMD
counterfactual baseline. Only in the 2000 and 2001 calendar year and only for NOx
emissions did RECLAIM facilities emit more pollution than what was expected under the
command-and-control alternative. In 2002 NOx emissions returned to a level below that
which would have been permitted under the command-and-control alternative. In all
years since the inception of the RECLAIM program SOx emissions have been below the
SCAQMD counterfactual. In each of these years – excepting 2000 and 2001 for NOx --
using this metric, RECLAIM outperformed the command-and-control alternative.
If the SCAQMD counterfactual is valid then it also follows that the total impact of
the program up until 1999 was a boon for regional air quality. Tables 4.1 and 4.2 show
the total number of unused RTCs for NOx and SOx in each year from 1994 – 2005.
These tables also show the percentage change of total emissions from year to year
compared to the first year the RECLAIM program was fully operational, 1994. As
previously noted, the total number of RTCs allotted is equal to the SCAQMD
counterfactual. Therefore, any RTC “left over” must represent an improvement over the
counterfactual alternative.

104
Table 4.1 Annual NOx Emissions for Compliance Years 1994 through 2005

Annual NOx
Emissions
(tons)
% Change
from 1994

Total
NOx RTCs
(tons)
NOx RTCs
Left Over
(tons)
NOx RTCs
Left Over
(%)
1994 25,314 0.0% 40,127 14,813 37%
1995 25,764 1.8% 36,031 10,267 28%
1996 24,796 -2.0% 32,017 7,221 23%
1997 21,786 -13.9% 27,919 6,133 22%
1998 20,982 -17.1% 24,678 3,696 15%
1999 20,775 -17.9% 21,013 238 1.1%
2000 20,491 -19.1% 17,197 -3,294 -19%
2001 15,721 -37.9% 15,693 -28 -0.18%
2002 10,943 -56.8% 14,044 3,101 22%
2003 9,942 -60.7% 12,484 2,542 20%
2004 9,953 -60.7% 12,477 2,524 20%
2005 9,556 -62.3% 12,484 2,928 23%

Table 4.2 Annual SOx Emissions for Compliance Years 1994 through 2005

Annual SOx
Emissions
(tons)
% Change
from 1994

Total
SOx RTCs
(tons)
SOx RTCs
Left Over
(tons)
SOx RTCs
Left Over
(%)
1994 7,232 0.0% 10,365 3,133 30%
1995 8,064 +11.5% 9,612 1,548 16%
1996 6,484 -10.3% 8,894 2,410 27%
1997 6,464 -10.6% 8,169 1,705 21%
1998 6,793 -6.1% 7,577 784 10%
1999 6,378 -11.8% 6,911 533 8%
2000 6,009 -16.9% 6,185 176 3%
2001 5,003 -30.8% 5,557 554 10%
2002 4,374 -39.5% 4,924 550 11%
2003 3,855 -46.7% 4,292 437 10%
2004 3,580 -50.5% 4,292 712 17%
2005 3,621 -49.9% 4,292 671 16%
Source: SCAQMD 2007

105
In virtually every year before 1999 aggregate pollution levels from RECLAIM
facilities were much lower than what the agency expected under command-and-control.
The early years of the program saw much lower NOx and SOx emissions than the
SCAQMD counterfactual baseline. 37% of NOx and 30% of SOx RTCs went unused in
1994. Thereafter these difference between allocation and use became smaller and
smaller. In 1999 only 1.1% of NOx RTCs and 8% of SOx RTCs went unused. As
figures 4.1 and 4.2 make clear, what accounts for this can mostly be attributed to a
reduction in the number of RTCs and not an increase in the amount of pollution emitted
from RECLAIM facilities. RTC allocations decreased from year to year until 2003, after
which they are predicted to be held nearly constant.
Obviously something dramatic happened in compliance years 2000 – 2001 which
have led many to reassess RECLAIM’s effectiveness. In 2000 RECLAIM NOx
emissions were 19% higher than the counterfactual baseline. In 2001 RECLAIM was
just slightly worse than expected with .18% more NOx emissions released than permitted.
Those figures are slightly misleading in that SCAQMD assessments are done for the
normal calendar year, but the time period for which an RTC is valid may overlap a
calendar year. In fact, in the first half of 1999 actual use of RTCs seemed to be on track
for being well below total allocation, but the second half of 1999 actually saw demand for
RTCs outstrip supply. If we were to bifurcate the year then we would come to the
conclusion that more NOx was released than permitted in the second half of 1999
(SCAQMD 2002). Between the second half of 1999 until the end of 2001, at least
judging by the SCAQMD counterfactual, RECLAIM had failed to perform as well as

106
command-and-control in delivering a reduction in NOx pollution. SOx emissions were
less under RECLAIM than the SCAQMD counterfactual in all years.
Early external assessments of RECLAIM also used the SCAQMD counterfactual
as a baseline for comparison. Many of these do not explicitly recognize the
counterfactual nature of RTC allotment as proxy for the subsumed command-and-control
regime, but nevertheless their analyses use it as a baseline for comparison. For example
Kamieniecki, Silver, and Shafie (1999, 112) using these same data come to the
conclusion that they, “suggest, therefore, that the RECLAIM program has been fairly
effective in reducing certain air pollutants.” Kerr, Anderson, and Jaksch (2000) came to
a similar conclusion and claimed, “RECLAIM is also meeting its environmental goals (p.
28)” and, “had succeeded in developing a functioning market” (p. 113). Later
assessments, especially those done during or shortly after 2000-2001 would not be so
kind.
Assessments of RECLAIM during the time period during or just after RTC use
was higher than RTC allotments vary in degree of sophistication but come to similar
conclusions: the program had failed to live up to expectations. Since these assessments
relied on the SCAQMD’s own counterfactual for comparison, how could they not? On
one end of the spectrum were journalistic assessments based on interest group input – on
the one side from some environmental groups that had never been keen on the idea of
markets for pollution and on the other side from large power producers who were the
program’s biggest early supporters but who had suddenly cooled to the idea of letting the
market decide how much an RTC should cost as the price skyrocketed – which came to

107
the conclusion that RECLAIM was a failure. One article in The Los Angeles Times is
indicative of the popular assessment:
It was supposed to be a revolutionary way to clean up the environment, a
business-friendly strategy to slash industrial emissions without the heavy hand of
government. But the Southland's market basket experiment has been a serious
disappointment (Polakovic 2001)

Beginning in 2001, this author also began a series of assessments of the
RECLAIM program using the SCAQMD counterfactual as proxy for what would have
happened under command-and-control and came to similar, if not more nuanced,
conclusions. For instance, in a 2002 paper I came to the conclusion that, “The claims that
market oriented regulations are more effective than command-and-control cannot be
verified by the RECLAIM experience” (Wright 2002, 37; see also Wright 2004; Wright
2006).
Perhaps the most scathing assessment came from an EPA sponsored evaluation
done in the wake of the widely reported market “failure”. This was done at a time when
the EPA was on the forefront of pushing cap-and-trade as a feasible solution to
environmental problems. The EPA explicitly asks and then answers the same question
raised here:
How have actual emission reductions compared to those that would have occurred
under the subsumed CAC system?

While there can be no definitive answer to this question, the data suggest that the
program has produced far less emission reductions than either were projected for
the program or could have been expected from the subsumed CAC system. (EPA
2002, 56-57)

108
The document leaves little doubt about what the EPA thought of the SCAQMD’s
handling of the RECLAIM cap-and-trade market. In their eyes RECLAIM was a failure.
It should be noted that the EPA study does take issue with the validity of the SCAQMD
counterfactual. Even so, they include it as part of a broader analysis since part of the
audience which the study was addressed was the SCAQMD itself which used the
counterfactual in internal assessments of the program. RECLAIM was underperforming
the command-and-control alternative as defined internally by the implementing agency’s
own measures and the EPA wished to remind them of it. But if RECLAIM was failing at
producing X amount of pollution reduction then implicit in the EPA criticism is that
command-and-control would have succeeded.
These analyses must be understood as having been written in the context of a
widely held belief that something had gone terribly wrong with the much heralded
RECLAIM program. Later assessments would rely less and less on the SCAQMD
counterfactual even as the SCAQMD defended it. Even after 2002, as NOx emissions
once again fell below those allowed under the allotted RTC permits, one can sense that
those following the program would no longer take the SCAQMD’s internal assessments
at face value (for instance see Zhan 2006; or Wright 2006). We will return to this in the
next section.
For the purposes of this study let us continue to assume that the SCAQMD
counterfactual is a valid baseline by which to compare RECLAIM against the command-
and-control alternative. Using this measure one would conclude that after 2002 the
RECLAIM program was once again outperforming the command-and-control alternative.

109
Returning to the data presented in figures 4.1 and 4.2 one can see that gross NOx and
SOx emissions under the RECLAIM program were less than the total number of RTCs
allotted. Between 2002 and 2005 SOx emissions varied from 11% to 17% lower than the
SCAQMD counterfactual. In those same years NOx emissions were 20% to 22% lower
than the SCAQMD counterfactual.
Even if one includes those years in which the SCAQMD counterfactual
outperformed RECLAIM in terms of NOx emissions, the gross amount of pollution under
the program was far less than what was expected under command-and-control. Using the
date shown in tables 4.1 and 4.2 one can calculate the net impact of the program’s life on
total emissions. Table 4.3 shows the total impact of RECLAIM on SOx emissions and
table 4.4 does the same for NOx emissions.

Table 4.3 Net impact of RECLAIM on SOx emissions 1994 - 2005
SOx Emissions

under 1994
baseline
(tons)
% Change
from 1994

Total SOx RTCs
Left Over
(tons)
Total SOx RTCs
Left Over
(%)
3,611 -50% 13,213 16%

Table 4.4 Net impact of RECLAIM on NOx emissions 1994 – 2005
NOx Emissions
under 1994
baseline
(tons)
% Change
from 1994

Total RTCs
Left Over
(tons)
Total NOx RTCs
Left Over
(%)
15,758 -62%
50,141 19%

110
From table 4.4 we cans see that over the life of the program 13,213 fewer tons of
SOx have been emitted under RECLAIM than the command-and-control alternative.
That is a 16% improvement over the SCAQMD counterfactual for SOx emissions. The
NOx improvement is even greater. Table 4.4 shows that RECLAIM fared 19% better
than the SCAQMD counterfactual, with an estimated savings of 50,141 tons of NOx
emitted over the 12 year period.
Using these metrics, the conclusion must be drawn that the cap-and-trade market
that is RECLAIM has outperformed the command-and-control AQMP that it replaced.
Had traditional command-and-control measures been in place, there would have been
more overall pollution. Even if command-and-control outperformed cap-and-trade from
late 1999 through 2001, the net effects of the program have still been positive. These
conclusions are not only my own, but are also reflected in the most recent publicly
available SCAQMD (2008) RECLAIM assessment which covers the 2006 calendar year:
aside from the effects of the California energy crisis on Compliance Years 2000
and 2001 emissions, it can be concluded that RECLAIM has achieved its targeted
emission reductions since aggregate emissions have been below aggregate
allocations during all other compliance years. (ch. 3, p. 1)

According to the agency charged with implementing the program and using their own
internal evaluative metric, RECLAIM has been an overall success.
There is an interesting side note as to the way in which the SCAQMD assesses the
RECLAIM program. In the wake of criticisms of the program –especially the scathing
EPA assessment – staff at the SCAQMD began to include another metric by which to
measure RECLAIM performance. This is reflected in later annual assessment, such as

111
the ones performed in 2007 and 2008, which compare yearly SOx and NOx output
against a starting year baseline. A decrease in emissions from the 1994 baseline is
presented as evidence that the RECLAIM program is working. These data were included
in the first column of charts 4.3 and 4.4.
If this is a valid metric, which the SCAQMD seems to believe, then in the year
2005 there were 15,758 fewer tons of NOx emitted than the 1994 baseline. That’s a 62%
reduction in NOx emissions from 1994 by facilities participating in the RECLAIM
program. SOx emissions are down 3,611 tons, a nearly 50% reduction. While not as
important a measure as comparing RECLAIM’s cap-and-trade system to a competing
command-and-control system, these data do show emissions trending in the hoped for
direction.

Problems with the SCAQMD counterfactual

One of the major criticisms of the SCAQMD counterfactual is that it does not
truly represent what would have happened under the command-and-control regime it
replaced. For instance, one objection raised by environmental groups in the early years
of the RECLAIM program was that the initial allocation of RTCs issued to facilities was
far too high (Thompson 2000). But participants argued that since the program would
begin during a recession year that the RTC baseline would be set too low if it was based
on the then (1993) current levels of emissions. Environmentalists argued that the supply
of RTCs was well above any conceivable demand schedule for the foreseeable future.

112
While command-and-control would require stationary facilities in the region to begin to
emit less SOx and NOx now, the RECLAIM program would not force these emissions
reductions until years later.
How valid were these objections? The method for allocating RTCs to affected
industries and therefore setting an initial baseline is specified in SCAQMD Rule 2002.
This rule specifies that RTCs will be issues based on the maximum emissions throughput
from each pollution source (not per firm, but per source meaning that each firm received
RTCs for each source under its control) from 1989 through 1992. This is the source of
many of the criticisms of the RECLAIM program since industries which were polluting
less at the time of adoption than they were in previous years would now be given license
to increase emissions. This would be the case regardless of whether or not those
reductions came as a result of a slack in use caused by the recession or from other factors
such as process changes, efficiency improvements, or the installation of emissions
reducing technologies such as “scrubbers”.
What is often overlooked is that Rule 2002 actually goes farther than that. The
initial formula for allocation also converts previously issued Emissions Reductions
Credits (ERCs) into RTCs. ERCs were offset credits given to industries in lieu of
reducing emissions at regulated facilities during the previous command-and-control
regime. For instance, oil refineries, stationary sources of pollution which would be under
the new RECLAIM regime, were allocated ERCs based on the number of old cars they
paid to have taken off the road and destroyed – even though these were mobile sources of
emission which would not be regulated under RECLAIM. The use of ERCs shows that

113
early on the SCAQMD was thinking outside the traditional command-and-control box
which usually proscribed best available technology requirements at the facility level.
Unlike a true cap-and-trade market, though, these ERCs could not be bought and sold on
the open market. Further still, non-RECLAIM participating facilities that had been
issued ERCs under the old command-and-control regime could also convert these to
RTCs and then sell them to RECLAIM participating firms. This inflated the number of
RTCs beyond even the maximum levels of pollution emitted in the pre-recession years.
All of these factors meant that in the initial years of the RECLAIM program firms
would not generally need to worry about emissions levels. RTC allocations were well
above the total amount of pollution which would be permitted. When supply is high and
demand is low, we would expect to find depressed prices. This would not necessarily
mean that the cap-and-trade had led to marginal reductions of pollution due to efficiency
gains – as predicted by economic theory and which is what the SCAQMD claimed (EPA
2002) – but could also indicate that there was simply too high of an initial allocation. In
other words, why try to lower emissions when no such lowering is required? This is
exactly what we find in RECLAIM.
The vast majority of RTC trades between 1994 and 1999 were done without price.
Figures 4.3 and 4.4 show RTC trades done with and without price for these years. They
also show the total amount spent on those trades. It should be remembered that RTCs are
allocated at the facility level and not the firm level. What this means is that a single large
firm may be allocated a certain number of RTCs for one facility and then another set of
RTCs for another facility. A power company, for instance, might be allocated multiple

114
sets of RTCs because it operates several different power plants across the region. Such a
firm would then be in the position of “trading” unused RTCs from one facility to another
internally and without price. They are free to increase pollution at one power plant as
long as another plant owned by the firm has extra RTCs.
This is not necessarily a problem since the goal is to lower total emissions and not
the emissions of every single facility. But if the initial allocation of RTCs was set so high
that few or no facilities actually lowered emissions at all, then it would be difficult to
claim that RECLAIM was making the kinds of inroads against pollution that would be
expected under command-and-control – unless, of course, one expected that the
command-and-control alternative would also produce no measurable reduction in the
targeted emissions.

Figure 4.3 Tons of NOx traded 1994 – 2004 with and without price

Source: SCAQMD 2005

115

Figure 4.4 Tons of SOx Traded 1994 – 2004 with and without price

Source: SCAQMD 2000

Figures 4.3 and 4.4 are consistent with the position taken by environmental
groups that far too many RTCs were initially allocated to RECLAIM participating
facilities. In the early years of the program few RTCs were actually traded on the open
market. While that trend continues, the price of RTCs in those early years also indicates
that there were more firms offering to sell RTCs than were willing to buy. Figures 4.5
and 4.6 represents the number of RTCs listed as “for sale” in the open market vs. how
many were actually bought. Because the chart was actually produced in the year 2000
and is from an assessment period two years prior to that. RTCs representing dates after
1998 should be treated as part of the futures market.

116

Figure 4.5 NOx RTC Supply and Demand (including futures)

Source: SCAQMD 2000

Figure 4.6 SOx RTC Supply and Demand (including futures)

Source: SCAQMD 2000

0
5000
10000
15000
20000
25000
30000
35000
1994 1996 1998 2000 2002 2004 2006 2008 2010
NOx RTC (Tons)
Year
RTCs Offered for Sale
RTCs Purchased by RECLAIM Facility
0
2000
4000
6000
8000
1994 1996 1998 2000 2002 2004 2006 2008 2010
SOx RTC (Tons)
Year
RTCs Offered for Sale
RTCs Purchased by RECLAIM Facility

117
These figures seem to confirm the earlier argument made by many in the
environmental community that RTC levels had been set too high. In all years more RTCs
were offered for sale at a price than were bought. Any “reduction” in the levels of RTC
permits over time did not actually represent a real reduction in SOx and NOx emissions
at facilities. Such “reductions” are what Darst (2001) terms “hot air” – pollution
reductions in name only and which may actually represent an increase in pollution.
When baselines used for comparative measurements are faulty, then the conclusions
drawn are also faulty. Early emissions “reductions” in RECLAIM are overstated and
probably represent such “hot air”.
The “unused RTC” figures touted by the SCAQMD representing RECLAIM
success might be interpreted in a different way. In the previous section we have seen
how the SCAQMD equated RTC use below allocation as a demonstration of equivalency
with command-and-control. The problem with this is that as soon as emissions exceeded
RTCs, then by definition this would mean that the program had failed to produce the
environmental benefits that would have occurred had RECLAIM not been adopted. If,
however, the initial allocation was so high that firms were not forced to lower emissions
at all, then the “unused RTC” figures are a useless metric.
All of this indicates that the SCAQMD counterfactual does not adequately
represent what would have happened had the former command-and-control AQMP been
kept in place. The argument that RTC levels are too high rests on the assumption that
command-and-control would have produced fewer emissions. It is argued that had
regulators followed the traditional path of insisting that firms adopt best-available-

118
technologies standards that pollution levels would have been reduced above what was
seen under RECLAIM. One way to test the validity of the counterfactual would be to see
how facilities regulated under command-and-control within the Los Angeles basin faired
compared to those regulated under the RECLAIM cap-and-trade program. Did facilities
that were regulated under command-and-control lower emissions at a greater rate than
facilities under RECLAIM?

A comparative approach: RECLAIM facilities vs. non-RECLAIM facilities

In order to compare emissions reductions between facilities under the command-
and-control regime and those under cap-and-trade, emissions data was acquired from the
California Air Resources Board (CARB). Stationary source facilities which are subject
to federal and state air quality regulations are tracked by the CARB which compiles &
publishes these firm level data (http://www.arb.ca.gov/html/databases.htm). These data
were downloaded in May of 2006 and include reporting years 1993 – 2001. The data
represent the entire universe of regulated facilities in the SCAQMD region. The
SCAQMD regulates pollution in all of Los Angeles and Orange counties as well as those
areas of San Bernardino and Riverside counties in close proximity to the Los Angeles
basin. They do not represent non-point or mobile sources of NOx and SOx emissions in
the region, such as mobile emissions sources like automobiles, which may be affected by
other regulations.

119
There are some gaps in the data. For instance no data is available for 1994.
Another striking problem is that many of the data points for an individual facility do not
change from year to year. There are two reasons for this. First, much of the emissions
data reported to CARB are based on an estimation process. This explains why there are
no differences between 1993 and 1995 data. It is difficult to believe that not a single
regulated facility that operated in both years neither increased nor decreased emissions
levels. Even if facility level fluctuations were minimal the net impact might be great.
Second, when there is a single year reporting gap CARB fills in missing data by
assuming that emissions from a facility were the same as the prior year. So, if a facility is
estimated to emit 60 tons of SOx in one year and no data is then reported for the
following year, then CARB fills in the missing data by assuming that the facility is still
emitting 60 tons of SOx. It is therefore difficult to assess whether or not data points
showing a facility emitting the same amount of pollution year after year is an accurate
reflection of a firm continuing to operate as it has in the past or if it simply reflects poor
reporting and enforcement practices.
Keeping these limitations in mind, these are the only data available which would
allow us to compare facilities operating under the RECLAIM cap-and-trade system with
facilities operating under traditional command-and-control. Comparing these two groups
approaches a controlled experiment (Gerring 2004). In the 1993 all facilities in the
region operated under the command-and-control AQMP giving us a baseline group. In
1994 the RECLAIM program was started dividing the universe into two groups: 1) The
“treatment” group, which consists of all facilities participating in the RECLAIM program

120
and which were thus subject to the policy intervention. 2) The “control” group, which
consists of all regulated facilities which were not subject to the “treatment”. Because the
latter group continues to be regulated under command-and-control, emissions levels from
these facilities should be an indication of how RECLAIM facilities would have
performed without the intervention.
Another benefit of comparing these two groups is that they all operate in the same
region and therefore are subject to the same regulatory agency, the SCAQMD. Many of
the variables that one might suspect could account for variation in emissions level are
then controlled for. For instance, Zhan (2006) controlled for such variables as economic
and population growth rates in several regression models he constructed trying to test
whether RECLAIM had any discernable impact on emissions levels. If one compares
only facilities within the greater Los Angeles region, though, presumably any socio-
economic trends which could be measured would affect all facilities equally. Also, any
differences in emissions levels which might be attributed to the regulatory agency are
controlled for since all are subject to the same regulatory agency’s jurisdiction.
The RECLAIM universe is very limited in number and does not regulate all
sources of NOx or SOx. When the program was started in October of 1993, only 394
stationary source facilities were included (SCAQMD 1994). Ten years after the program
was first implemented, 332 facilities were included (SCAQMD 2003). This number
includes 94 facilities that ceased operation, 67 which have been excluded from the
program, and 99 facilities which opted in. The RECLAIM “bubble” then consists of only
those facilities which have been issued RTCs. These facilities are all stationary sources

121
of NOx or SOx. In 1994 facilities in the RECLAIM universe accounted for only 42% of
NOx emissions from stationary sources in the region. The majority of NOx in the greater
Los Angeles basin come from non-stationary sources such as automobiles. In 1994
RECLAIM facilities only accounted for 11.3% of all emissions in the region (Kerr 2000).
Table 4.5 indicates the number of facilities looked at in this section of the study.
It should be noted that these data represent the entire universe of regulated stationary
source facilities within the SCAQMD region. Facilities which either began operation or
ceased to operate during this time period were excluded from the analysis. Thus the
analysis only includes facilities which were already in operation at the time of
RECLAIM’s adoption and which were in continuous operation during those years. This
controls for the net affects of plant closures, which could be great. If some facilities in
either the RECLAIM or non-RECLAIM group are shut down then presumably net
pollution levels will also decrease. There is no way of knowing if such a decrease due to
plant closures could be attributed to the policy intervention since plant closures are
common in all industries. When a facility in either group ceased to emit for a number of
years but later began to emit again, it was included for analysis. This is because under a
cap-and-trade program a firm operating more than one facility might find that the least
costly way to comply with the RTC cap is to shut down a less efficient facility and
operate other more efficient facilities at greater capacity. If, however, technological
breakthroughs were to be found at some point of time in the future which were cost
effective, then the “closed” facility might be reopened. We can then compare how

122
facilities under command-and-control fared vs. those under the cap-and-trade regime over
the same period of time.

Table 4.5 RECLAIM vs. Non-RECLAIM universe
RECLAIM
Universe NOx
Non-RECLAIM
Universe NOx
RECLAIM
Universe SOx
Non-RECLAIM
Universe SOx
Total N
232 facilities 915 facilities 109 facilities 147 1363

Unlike the SCAQMD counterfactual, the data presented in tables 4.6 – 4.9 that
follow are actual emissions data and not a simple accounting of unused RTCs. The
difference may seem subtle to those not familiar with the way RTCs are generated but
they are important. The SCAQMD has long allowed RECLAIM firms to offset RTCs by
reducing emissions outside the “bubble” in cap-and-trade. The most important of these
schemes allowed RTCs to be “created” by the scrapping of old cars. Instead of actually
lowering emissions at a facility, firms were given the option of lowering emissions on
non-RECLAIM sources which would then be counted as an offset against actual
emissions in the form of a newly created RTC. Thus a facility that was able to exercise
this option might actually lower the total level of NOx in the affected region while
actually polluting more at the facility level. While this may lead to overall environmental
improvements, it does nothing to inform us about facility level behavior under cap-and-
trade. A truly valid measure of firm level behavior should exclude environmental
improvements generated outside what would normally be thought of as the “bubble” in a
cap-and-trade system. Especially given the SCAQMD’s history of also letting firms
regulated under command-and-control also offset emissions.

123
Table 4.6 - 4.9 shows gross emissions for the two targeted pollutants for the entire
universe of both RECLAIM and non-RECLAIM facilities in continuous operation
between 1993 and 2001. They also show the total change in emissions by ton from year
to year. That is, how much more or less these facilities are emitting in one year from the
previous year. They also compare year to year changes from a 1993 baseline in tons.
1993 was chosen because it was the last year in which all facilities were regulated under
the previous command-and-control regime. By whatever baseline one wishes to use, one
would hopefully find that emissions levels are going down from year to year. Even if
there is a temporary increases in emissions for a singly year, one would hope that overall
emissions would be lower than in the baseline year.
These tables also express changes in emissions levels as a percentage of the
previous year and as a percentage change from the 1993 baseline. While interesting to
note the total impact a program may have on overall emissions of a pollutant, what we
are really interested in knowing here are the relative impacts of two separate regulatory
regimes. With a common denominator we can compare the percent of emissions reduced
for RECLAIM facilities vs. those reduced at non-RECLAIM facilities. The underlying
assumption is that had RECLAIM facilities continued to have been regulated under
command-and-control then they would have behaved in a similar fashion to those found
in the non-RECLAIM universe presented in the data that follows.

124
Table 4.6 NOx Emissions for continuously operating RECLAIM facilities
RECLAIM (tons)
Yearly Change
(tons)
Yearly Change
(%)
Change
from 1993
(tons)
Change from
1993 (%)
1993 21978.2 0 0 0 0
1995 21978.2 0 0 0 0
1996 22341.784 363.584 1.6542938 363.584 1.6542938
1997 18355.165 -3986.619 -17.84378096 -3623.035 -16.48468*
1998 18355.664 0.499 0.002718581 -3622.536 -16.48241
1999 17956.819 -398.845 -2.172871545 -4021.381 -18.29714
2000 18355.165 398.346 2.218355044 -3623.035 -16.48468
2001 13945.011 -4410.154 -24.02677394 -8033.189 -36.55071

Table 4.7 NOx Emissions for continuously operating non-RECLAIM facilities

NOX Non
RECLAIM (tons)
Yearly Change
(tons)
Yearly Change
(%)
Change
from 1993
(tons)
Change from
1993 (%)
1993 4899.4 0 0 0 0
1995 4899.4 0 0 0 0
1996 5109.988 210.588 4.298240601 210.588 4.298240601
1997 3882.683 -1227.305 -24.01776677 -1016.717 -20.7518675*
1998 3882.923 0.24 0.006181293 -1016.477 -20.74696902
1999 3882.683 -0.24 -0.006180911 -1016.717 -20.75186758
2000 3882.683 0 0 -1016.717 -20.75186758
2001 3701.772 -180.911 -4.659432666 -1197.628 -24.44438094

Table 4.8 SOx Emissions for continuously operating RECLAIM facilities
Non-RECLAIM
(tons)
Yearly Change
(tons)
Yearly Change
(%)
Change from
1993 (tons)
Change from
1993 (%)
1993 6195.500 0 0 0 0
1995 6195.500 0 0 0 0
1996 7835.253 1639.753 26.467 1639.753 26.467
1997 6182.947 -1652.306 -21.088 -12.553 -0.203
1998 6182.798 -0.149 -0.002 -12.702 -0.205*
1999 7197.014 1014.216 16.404 1001.514 16.165
2000 6182.947 -1014.067 -14.090 -12.553 -0.203
2001 6554.773 371.826 6.014 359.273 5.799

Table 4.9 SOx Emissions for continuously operating Non-RECLAIM facilities
Non-RECLAIM
(tons)
Yearly Change
(tons)
Yearly Change
(%)
Change from
1993 (tons)
Change from
1993 (%)
1993 314.200 0 0 0 0
1995 314.200 0 0 0 0
1996 397.408 83.208 26.48249523 83.208 26.4824952
1997 296.259 -101.149 -25.45218013 -17.941 -5.7100573
1998 296.259 0.000 0 -17.941 -5.7100573*
1999 296.259 0.000 0 -17.941 -5.7100573
2000 296.259 0.000 0 -17.941 -5.7100573
2001 321.201 24.942 8.41898474 7.001 2.2281986
* indicates a t-test of statistical significance at the .05 level

125

How do facilities regulated by traditional command-and-control fare compared to
those under the RECLAIM cap-and-trade program? The results are surprising and were
not what this author predicted. Some of these data also seem to fly in the face of much of
the criticisms leveled at RECLAIM. Figure 4.7 is a visual representation of the gross
level of NOx emissions for RECLAIM and non-RECLAIM facilities in the SCAQMD
managed areas of Southern California. Figure 4.8 represents the same data, but for SOx
emissions.
NOx Emissions for RECLAIM facilities did fall after the adoption of the cap-and-
trade market. In 1993 RECLAIM facilities emitted 21978 tons of NOx. In 1996 yearly
emissions increased slightly but began to fall below 1993 levels by 1997. By the end of
2001, in the middle of the California energy crisis and in the same year that more RTCs
were used than were allocated, the net output of NOx from RECLAIM facilities had
actually fallen from the 1993 baseline by 8033 tons.
In those same nine years non-RECLAIM facilities also emitted less NOx. At the
start of 1993 they collectively emitted 4899 tons of NOx. In the following years there is
a steady downward trend interrupted only in 1996 when facilities emitted about 211 tons
more than in previous year and in 1998 when they emitted less than one ton more. At the
end of 2001 they emitted 3702 tons, a decrease of 1198 tons of NOx.
Trends in the emissions of SOx are a little less straightforward. In 1993
RECLAIM emitted 6195 tons of SOx. By 2001 they were actually emitting 359 more
tons of SOx than they were in the baseline year. In the years between SOx emissions

126
fluctuated both over and under the baseline. But non-RECLAIM facilities were also
emitting 7 tons of SOx more in 2001 than they were in 1993.

Figure 4.7 Gross NOx Emissions: RECLAIM vs. Non RECLAIM facilities

0
5000
10000
15000
20000
25000
1993 1995 1996 1997 1998 1999 2000 2001
NOX RECLAIM
NOX Non RECLAIM

127
Figure 4.8 Gross SOx Emissions: RECLAIM vs. Non-RECLAIM facilities

But as already noted, the most important findings are the relative impacts of
participation in the RECLAIM cap-and-trade market. Figure 4.9 illustrates yearly
changes in SOx emissions for RECLAIM and non-RECLAIM facilities as expressed as a
percentage change from the previous year. It visually captures fluctuations in emissions
outputs by facilities from 1993 to 2001. As can be seen, while there was considerable
variation in year to year in the amount of SOx emitted by regulated facilities, those
variations tend to move in the same direction. So, that while RECLAIM facilities emitted
26% more SOx in 1996 than they had in 1995, non-RECLAIM facilities also emitted
about 26% more in the same year. And in 1997 when those in the RECLAIM universe
emitted 21% less than the previous year, non-RECLAIM facilities emitted 25% less. In
1999 and 2000 there was much more variation from the mean for RECLAIM facilities
0.000
1000.000
2000.000
3000.000
4000.000
5000.000
6000.000
7000.000
8000.000
9000.000
1993
1995
1996
1997
1998
1999
2000
2001
SOX RECLAIM
SOX non RECLAIM

128
than for non-RECLAIM facilities which held constant during these years. However, a t-
test showed that these variations were not statistically significant at the .05 level.
By the end of 2001 RECLAIM facilities were emitting almost 6% more SOx than
they were in 1993. This is well above the figure touted by the SCAQMD for that year
which reported that 10% of RTCs went unused. RECLAIM facilities that had been in
operation since the beginning of the program were actually polluting more under the cap-
and-trade market than they were under command-and-control. However, non-
RECLAIM facilities still operating under the command-and-control rules were also
emitting more SOx. In 2001 the non-RECLAIM universe emitted 2% more SOx than
they had in the 1993 baseline year. This indicates that while command-and-control had
faired poorly at reducing this pollution, the cap-and-trade system had done even worse.
The only statically significant difference between the RECLAIM and non-RECLAIM
group was in the percent change in 1998 from the 1993 baseline. While statistically
significant, the actual difference between the two groups was only about 5%.

129
Figure 4.9 Changes in SOx emissions for RECLAIM vs. Non-RECLAIM in %

Figure 4.10 Changes in SOx Emission from 1993 baseline

-30
-20
-10
0
10
20
30
1993 1995 1996 1997 1998 1999 2000 2001
Non-RECLAIM Yearly
Change (%)
RECLAIM Yearly
Change (%)
-10
-5
0
5
10
15
20
25
30
1993
1995
1996
1997
1998
1999
2000
2001
Non-RECLAIM
Change from 1993 (%)
RECLAIM Change
from 1993 (%)

130
Similar trends can be seen in NOx emissions variations as measured as percentage
changes. While the degree of yearly variation for RECLAIM and non-RECLAIM
facilities differs, the direction of those changes from year to year is generally in the same
direction. When RECLAIM facilities tended to emit less NOx, so too did non-
RECLAIM facilities. For instance, in 1996 the RECLAIM universe emitted 1.65% more
NOx than it had in the previous year. But the non-RECLAIM universe also emitted more
and at a greater rate than their cap-and-trade counterparts: 4.3% more. In 1997 the
command-and-control universe outperformed RECLAIM by emitting 24% less than in
1996. The RECLAIM universe also dramatically improved, but only by 17%. Between
1998 and 2000 emissions levels for the non-RECLAIM universe held constant while
there was slight fluctuation up and down for RECLAIM facilities. In 2001 there was a
dramatic drop in RECLAIM NOx emissions and a less dramatic drop in non-RECLAIM
facilities. In that year the cap-and-trade universe emitted 24% less NOx than it had in the
previous year. The command-and-control facilities emitted less, but by only 4.7%. Since
much of the variation between the two groups can be explained by outliers, what appears
to be at least one year where there was a dramatic difference (2001) is not statistically
significant.

131
Figure 4.11 Changes in NOx emissions for RECLAIM vs. Non-RECLAIM in %

The most dramatic improvement in pollution is illustrated in figure 4.12. As one
can see, only in 1996 were NOx emissions by facilities in continuous operation ever
higher than the 1993 baseline. This is true of both RECLAIM and non-RECLAIM
facilities. In 1997 there is a rapid decline in emissions for both groups, but the decline in
the command-and-control group is slightly greater than for those under cap-and-trade: a
16.5% decline for RECLIAM and nearly a 21% decline for non-RECLAIM. There is a
leveling off in universes from that time until 2000 after which there is another rapid
decline in emissions for both groups. This time, however, the RECLAIM universe
outperformed the non-RECLAIM universe. By the end of 2001 the RECLAIM facilities
were emitting nearly 37% less NOx than they had in 1993. Non-RECLAIM facilities
were emitting 24% less. The difference between the two was not statistically significant.

-30
-25
-20
-15
-10
-5
0
5
10
1993 1995 1996 1997 1998 1999 2000 2001
Yearly Change
RECLAIM(%)
Yearly Change Non-
RECLAIM (%)

132
Figure 4.12 Change in NOx emission from 1993 baseline

Compare that to the SCAQMD report that RECLAIM facilities had emitted 0.18% more
than RTC allocations in 2001. To them and to others this represented a failure of the cap-
and-trade program at delivering similar results as command-and-control. But if these
data are correct then the problem is not with RECLAIM facilities emitting more than they
would have under command-and-control, the problem is with the assumption that the
counterfactual was a valid measure.

Conclusions

Did RECLAIM meet its effectiveness goals? Mostly. It really depends on what
one means by effectiveness and how one measures these goals. This chapter has shown
that measuring policy success and policy failure is dependent upon measurement
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
1993 1995 1996 1997 1998 1999 2000 2001
RECLAIM Change
from 1993 (%)
Non-RECLAIM
Change from 1993
(%)

133
standards. Those who designed the RECLAIM cap-and-trade market imagined that the
command-and-control regulations which it replaced would follow a consistent and
predictable downward emissions path. By the SCAQMD’s own standards, the
RECLAIM program has been a mixed bag. On the one hand the number of RTCs used
has nearly always been below that of RTCs issued and this represents overall success of
the cap-and-trade market in reducing emissions comparable to what they expected would
have occurred under command-and-control. On the other hand, the dramatic price spikes
for RTCs in 2000 – 2001 came as a shock to nearly all involved; and in those years
RECLAIM facilities emitted more NOx than the SCAQMD counterfactual. This
represented a temporary failure for the SCAQMD and seemed to vindicate many of the
concerns environmental advocates had over the initial allocation of RTC credits. One
assumption the environmental movement shared with the SCAQMD was that command-
and-control would have led to a predictable downward trend in emissions from regulated
facilities. But a closer look through a comparative analysis of how facilities in the
region that continued to be regulated under command-and-control show that their
emissions trends were no more predictable than the path of taken by cap-and-trade.
RECLAIM had not performed as expected, but neither had command-and-control. Ex-
ante predictions about what should happen are no substitute for ex-post analysis of what
actually happened.
RECLAIM had mostly met statutorily set goals, but those statutorily set goals did
not actually represent the path of command-and-control that these goals are claimed to
represent. Emissions of NOx had declined, but SOx emissions had not. However, the

134
SOx emissions of non-RECLAIM facilities had also not declined over the same period.
What is clear from these data are that RECLAIM did not produce any additional
environmental improvements over and above what would have been realized under
command-and-control. Thus, the first benefit attributed to cap-and-trade was not seen in
the RECLAIM program.
Why is this? The next chapter explores how RECLAIM lived up to the two other
expectations predicted of cap-and-trade: efficiency and innovation. As argued in chapters
two and three, efficiency and innovation are causal drivers in a cap-and-trade program
which should lead to greater environmental benefits. Assessing the extent to which
RECLAIM is efficient and has led to innovation can help explain why we have not seen
emissions reductions beyond command-and-control. An implementation framework is
used to explain how cap-and-trade is translated from idea to operating policy.

135
Chapter 5 Efficiency and Innovation
Command-and-control and cap-and-trade programs both target specific
populations in order to get them to reduce emissions which harm the environment or
human health. Under command-and-control regulators proscribe the way in which
emissions will be reduced, without setting an overall limit on those emissions. Under
cap-and-trade regulators set an overall limit on emissions without proscribing the way in
which those emissions will be reduced. In theory, both types of programs will reduce
overall emissions of a pollutant. Cap-and-trade is said to have several benefits not
captured by its command-and-control alternative. First, they may not only meet goals
like command-and-control but may also exceed them. Second, they are more efficient
than command-and-control. Third, unlike command-and-control they should force
innovation.
The previous chapter focused on the first benefit. It looked at a single aspect of
programmatic efficacy, relative impact; which asks the question how much is this
program affecting that outcome compared to the next most likely regulatory alternative.
As shown in the previous chapter, measuring a program’s relative impact may not be as
easy as a cursory glance might suggest. If one measures the relative impact of this
particular cap-and-trade program by comparing it strictly against program goals as
operationalized by the implementing agency in the number of RTCs issued then one
would come to the conclusion that the relative impact of RECLAIM is fairly high. In
most years and for the two pollutants regulated, the program has successfully met or
exceeded its goals. However, as shown in the previous chapter, the baseline of RTCs

136
used by the implementing agency probably does not accurately reflect the stated goal of
the program, which is to reduce emissions by as much as the command-and-control
regulations which it replaced. By comparing emissions levels of facilities targeted by
RECLAIM to those which are not, evidence was presented that the program did no better
than the command-and-control alternative. Hence, the first benefit of a cap-and-trade
market—that it has the potential to reduce emissions beyond what could be expected
under command-and-control—was not found in the RECLAIM program. Since a first
glance of RECLAIM shows that it is larger and more heterogeneous than other cap-and-
trade programs, such as Phase I of the Title IV program, RECLAIM seemed to have all
the right ingredients for successfully reducing emissions beyond what would have been
seen under command-and-control.
In this chapter I reiterate the findings presented in chapter four, and then make a
preliminary inquiry at the extent to RECLAIM shows the two other benefits said to come
from cap-and-trade markets. The literature on implementation and property rights,
explored in chapter two of this work, are used to explain why RECLAIM did not produce
as efficient a market as it could have, and why the program has not led to entrepreneurial
innovation. Efficiency and innovation are important because not only are they goals of
cap-and-trade but also because they help explain why the RECLAIM program failed to
produce environmental benefits above and beyond what would have been realized under
command-and-control. Looking at these two goals should help illuminate the findings in
the previous chapter. As viewed through an implementation framework, one can better

137
understand why the command-and-control ideal was not realized in the RECLAIM
program.
Figure 5.1 shows the same information found in Figure 1.3 found in chapter one
of this dissertation, but adds the findings from chapter four and from what follows in this
chapter. It lists the possible benefits of cap-and-trade compared to command-and-control
as predicted by economic theory and how the RECLAIM program actually performed.
The section which follows is a discussion of the information presented in it.

138
Table 5.1 CAC & CAT expectations vs. RECLAIM findings
Command-and-Control Cap-and-Trade
(Theory)
Cap-and-Trade
(RECLAIM)
Goal Attainment Yes. Goals defined by
technology known to
regulators. Goals
adjusted up or down in
light of technological
improvement.
Yes + possibility of
exceeding goals.
Mostly. Most years
exceeded goals but
under years of stress on
the system program did
not meet expectations.
Alternative measures
also show that SOx
emissions increased
rather than decreased,
but true of RECLAIM
and CAC facilities.
Efficiency No. Yes. Efficiency gains
predicted based on
differences in marginal
costs of abatement
between facilities.
Not clear. Because of
initially high RTC
permitting early low
prices not meaningful
measure. High prices
during stress years may
indicate less efficient
than hoped for. Lower
prices after stress years
may reflect rules
changes and not
program efficiency.
Innovation No. Reliance on
regulators to find best
available technology.
Yes. Technology
forcing as firms try to
find lowest cost way to
achieve goals inspiring
entrepreneurial activity.
No. Initial high RTC
permitting and therefore
low prices meant little
incentive to find new
technology. When
prices finally rose to
level which could force
technological
innovation, regulators
abandoned system
leading to lower prices
and reinstituted some
command-and-control
fixes for big polluters.

139
Goal attainment

1) Did RECLAIM meet its goals? Mostly.
As the data presented in chapter four shows, RECLAIM facilities emitted less
SOx than allowed in all years from the program’s inception. However, the alternative
measurement outlined in that chapter also show that SOx facilities actually increased
emissions rather than decreasing them in the time period under examination. Non-
RECLAIM command-and-control facilities also increased in SOx emissions as well.
Legally these goals were met, but if the overarching goal of the program was to reduce
SOx emissions then that did not happen. The alternative measure of command-and-
control also shows that these facilities also were emitting more eight years after the
AQMP was adopted. If RECLAIM had failed to meet its overarching goal, then so too
had the command-and-control alternative.
RECLAIM facilities also emitted less NOx than legally allowed in all years
except two. But the cumulative effect of the program was that, even counting the two
years when NOx goals were not achieved, net pollution levels were significantly lower
than net goals. The alternative measurement proposed in chapter four shows that NOx
emissions were down considerably from the beginning of the program.

2) Did RECLAIM exceed its goals? No.
Economic theory suggests that the possibility exists that a cap-and-trade market
could lead to environmental improvements above and beyond those found in command-

140
and-control. As the data presented in chapter four shows, officials at the SCAQMD have
consistently claimed that RTC levels represent a command-and-control counterfactual –
the amount of emissions that would have occurred under the subsumed command-and-
control regulations. If that was the case, then except for the two years in which NOx
emissions were higher than permitted, then we would have to answer that RECLAIM did
exceed its goals. However, as chapter four also showed, the SCAQMD counterfactual
did not accurately predict the emissions path of command-and-control facilities. As the
data make clear, RECLAIM facilities did not significantly differ in their emissions
patterns than their non-RECLAIM counterparts. At times RECLAIM facilities emitted
more than the command-and-control group, and at times they emitted less; but rarely
were these differences statistically significant, and when they were, the impact of those
differences was only slight. RECLAIM did not produce significant emissions reductions
over those found under command-and-control.

Efficiency

1) Is RECLAIM efficient? Not clear.
As discussed throughout this dissertation, efficiency can be measured in a number
of different ways and is both a goal of a cap-and-trade program—a dependent variable—
and a causal driver serving the overarching goal of emissions reductions—an independent
variable. While this dissertation does not attempt to thoroughly measure the efficiency of
RECLAIM, the efficiency of the program has been called into question throughout this

141
work. In the previous chapter the data showed that especially problematic to the
program’s efficiency were: 1) the initially high number of RTCs; 2) the generation of
RTCs external to the “bubble”, or targeted universe; 3) rules changes meant to lower the
price of RTCs. In light of these data, this section examines – in a cursory way – aspects
of the efficiency of the RECLAIM program as outlined in chapter three.

1) Market efficiency.
According to Hahn and Noll (1982), at least three conditions are necessary for
markets to work efficiently. Kamieniecki, Shafie, and Silvers (1999, 113-114) adopt
these conditions in their study of RECLAIM:
1a) “There must be an adequate number of participants in the market”. Nearly
all observers of the RECLAIM program claim that sufficiently high numbers of
participants exist in order for the market to be efficient (EPA 2002; Kaminiecki, Shafie,
and Silvers 1999; Kerr, Anderson, and Jaksch 2000; Johnson and Pekelney 1996; OECD
2002). As discussed in chapter three, though, it is not clear how any of these observers
come to this conclusion. There may very well be a sufficient number of participants in
the market, but it is not clear below what number insufficiency occurs. Clearly, though,
RECLAIM is one of the largest cap-and-trade markets in existence. Phase I of the Title
IV SOX program, which has been unanimously seen as an efficient market, included only
263 of the dirtiest large coal fired power plants in the country (Ellerman et al 2000, 6). If
263 is the baseline for market efficiency, then the 400+ facilities participating in the

142
RECLAIM program in any given year must seem to meet this condition to most
observers.
But the way in which the program has actually been implemented is far more
complex. Implementation studies have often show how an ideal is changed as it meets
the complexities of the real world (Derthick 1972; Pressman and Wildavsky 1973).
Bardach 1977). The true number of participants in the RECLAIM market is much
smaller. Little attention has been given to the complicated nature of RTCs and the rules
governing their trades in the RECLAIM program. RECLAIM RTCs are subdivided in a
number of different ways. First, the RECLAIM program actually created two markets:
one NOx market and one SOx market. Thus, each market only has a fraction of the
overall 400 facilities in it.
Another aspect of the program which has merited virtually no discussion is that
each market is actually subdivided even further. Each facility is geographically listed as
being either in Zone 1 or Zone 2, reflecting facility proximity to the coast. A facility in
Zone 1 may buy RTCs only from other facilities Zone 1, but a facility in Zone 2 may buy
RTCs from either zone. This reflects equity concerns on the part of the SCAQMD about
the impact of increasing pollution at some facilities along the coastal Zone 1. The
majority of emitting facilities are located in the coastal zone. Prevailing winds from the
Pacific Ocean blow pollution inland over the course of the day. Regulators believe that
by dividing the market into zones, they will reduce further transference of emissions.
This geographic division also reduces the likelihood that “hot spots” will occur. While
the division may be based on valid concerns, the effect of limiting trades by geography is

143
to limit the number of RTCs available for trade on the open market. A facility in Zone 1
is much more restricted in the RTCs they can purchase.
RTCs are further divided into seasonal cycles. Cycle 1 runs the normal calendar
year and Cycle 2 runs from June of one year to July of the next. What at first glance
seems to be a very large and therefore presumably robust market is thus divided by
pollutant, by region, and then by a yearly cycle. But the most troubling division of the
market came in the wake of the California power crisis in 2000-2001. In those years the
RECLAIM NOx market was further divided into power production facilities and non-
power producing facilities. The justification of this division was that owing to the
increased demand for power during the crisis, power companies were bidding up the
price of RTCs on the open market causing a spike in prices. Thus, for those years at
least, the market was divided once again.
The totality of these trading restrictions show that the market as implemented was
much smaller than the theoretical market often discussed in the literature and is consistent
with implementation studies showing that conflicting or inconsistent goals affect policy
outcomes (Bullock 1981; Calista 1994; Pressman and Wildavsky 1973; Mazmanian and
Sabatier 1983). In the case of RECLAIM a conflict exists between the minimum
numbers of participants needed to create market efficiency and the unique geographic
context of the Los Angeles basin where prevailing winds blow air pollutants inland
throughout most of the day, leading to the bifurcating of the market into Zones. LA’s
geographical context also explains why RTCs were further divided into cycles. Much
ground level ozone (“smog”) in the area is caused by a chemical reaction of NOx and

144
other pollutants with sunlight. Thus, smog is worse in the summer than it is in the winter.
By temporally bifurcating RTC trades the SCAQMD ensures that emissions won’t be
pushed from winter months to summer months where the effects of those emissions
would impact the environment and human health more adversely. Potential efficiency in
the market is traded for other competing values.
Despite these many restrictions, it remains unclear whether the RECLAIM market
is large enough to promote efficiency. Much further study is needed into the minimum
number of participants needed to satisfy this condition. What is clear, though, is that the
number of participants in each of RECLAIM’s two markets is much smaller than many
who look at the program have noted. The market efficiency of RECLAIM is therefore
probably overstated.
1b) “Transaction costs must be sufficiently low”. The most direct form of
transaction costs in cap-and-trade markets come from regulatory oversight. One widely
heralded aspect of the RECLAIM program has been that no prior regulatory review is
required of RTC trades. Another key to keeping transaction costs low is the emergence
of transparent trading mechanism where buyers and sellers can easily attain price
information. While transparency has generally not been considered a problem in
RECLAIM since all trades and their clearing prices have always been reported via the
SCAQMD website (Kamieniecki, Shafie, and Silvers 1999), recent rules changes indicate
that the SCAQMD and market participants believed that the system was not transparent
enough in its early years leading to higher transaction costs.

145
The original rules did not have a timely filing requirement. What this means is
while all trades had to be reported, when they were reported was left up to each party in
the transaction. In 2001 one rule change included requirements for reporting RTC trades
within five business days of trade. The SCAQMD also began posting this information
online in a timelier manner, increasing the transparency of the trading mechanism
(SCAQMD 2007; see also Colby 2000).
Another change has been the inclusion of provisions for “infinite year RTC” (IY
RTC) trading (SCAQMD 2007). One of the key complaints of firms targeted by the
program has been that no “banking” of RTCs is allowed. The Title IV SOx program
allows unused pollution credits to be “banked” and then used at a later date. But this
reflects the cumulative nature of acid rain on the environment over the entire continental
United States. In the Los Angeles basin, and especially in the case of NOx, the health
affects of the regulated pollutant are immediate. More NOx pollution today means more
ground level ozone today. So banking has never been an option for regulated firms.
However, an efficiently running futures market serves the vital function of firms being
able to plan for future fluctuations in emissions levels. Under the new trading rules,
RTCs from one facility can be permanently transferred to another. This decreases costs
as only one transaction is needed to move RTCs between facilities rather than multiple
transactions of yearly futures contracts.
These rule changes show that while transaction costs have never been especially
high in the RECLAIM market, that there was room for improvement over the lifetime of
the markets implementation. Further, adjustment of the program by implementing agents

146
did not always have a negative affect on the market. These rules changes suggest that the
implementing agents and targets had gone through some amount of learning (Sabatier and
Jenkins Smith 1993) and that the implementing agency is committed to the policy’s
overarching goals (Hjern and Hull 1983).
1c) “Government must provide sufficient monitoring and enforcement.”
Nearly 80 percent of emissions under RECLAIM are from major sources, which are
monitored by Continuous Emissions Monitoring Systems (CEMS). CEMS equipment
can collect emissions data in nearly real time. These data are generally reported to the
SCAQMD on a daily basis. The other 20 percent of facilities have equipment installed
which calculates emissions either weekly, for medium sized facilities, or monthly, for
smaller facilities. Regardless of facility size, the SCAQMD does yearly audits of
emissions equipment to ensure the reliability of the data generated by them. Over the
lifetime of the program auditing shows increased reliability from CEMS equipment and
the diffusion of this equipment from larger facilities to smaller ones (SCAQMD 2007).
This means that emissions data generated in more recent years is more reliable than data
generated at the outset of the program.
There have been occasional problems with monitoring, the most common of
which is missing data, but a content analysis of 12 years of SCAQMD reports shows that
these monitoring problems have been minimal, usually with less than a dozen facilities
found to have monitoring errors in a single year. When missing data is found, the
SCAQMD also has procedures by which to calculate the missing error values (SCAQMD
2007).

147
Enforcement has not been a major problem either. The SCAQMD types
RECLAIM violations into five separate categories: allocation, monitoring,
recordkeeping, reporting; and operational. The majority of enforcement actions have
been handled administratively and were generally due to minor reporting and monitoring
errors. A major exception to this came during the California power crisis when the
governor declared a “state of emergency” and temporarily suspended power companies
from the requirements of the California Clean Air Act. Because it was this act that
legislatively enabled the RECLAIM program, two RECLAIM participants — the Los
Angeles Department of Water and Power and AES Alamitos, LLC, both major electrical
generators in Southern California – believed that this and related actions taken by other
regulating authorities relieved them of their compliance responsibilities. In 2000, the
SCAQMD was informed that both sources intended to operate as necessary to meet
energy demand and that they expected to pay no enforcement penalties (SCAQMD
2007). The SCAQMD disagreed with this interpretation of the governor’s actions. These
two facilities alone accounted for 70% of NOx emissions over RTC allocation in 2000-
2001. The Los Angeles Department of Water and Power eventually agreed to fund
environmental projects to the tune of $14 million dollars in lieu of compliance but did so
before the full extent of the violations were known and in anticipation of them. AES
Alamitos eventually paid a $17 million dollar cash fine to the SCAQMD.
In other words, they believed that it was cheaper to pay the negotiated
enforcement penalty than it was to buy additional RTCs to cover excess emissions on the
open market. By February of 2001 NOx RTC prices had reached their peak. The

148
highest recorded price was $124,000 per ton. At this point it was cheaper to pay fines
and penalties than to install new emissions equipment or take other actions to reduce
emissions.
Thus there is no evidence in RECLAIM of a common finding in the
implementation literature of agents not fully embracing program goals and thus, through
enforcement inaction, thwarting the intent of policymakers (Pressman and Wildavsky
1973; Bardach 1977). The rules establishing the policy are fairly clear and
straightforward and offer little wiggle room for enforcement (Mazmanian and Sabatier
1983). In the case of RECLAIM, monitoring and enforcement have not been a major
issue. Only occasionally, and in the context of the energy crisis, have there been
purposeful violations of the rules by targeted facilities. As seen in chapter three, the
major concern in the literature here has been that regulators would lack the tools
necessary to monitor or would lack the will to enforce. The problem with enforcement
during the power crisis had nothing to do with either of these factors. The problem was
that the program design had never anticipated a situation in which RTC prices would ever
reach a sufficient level that some participants found that it was cheaper to pay
enforcement penalties than it was to comply.

Credibility and commitment to property rights

In addition to the three aspects of market efficiency first proposed by Hahn and
Noll (1982) and adopted by Kamieniecki, Shafie, and Silvers (1999) listed above, Glazer

149
and Rothenberg (2001) add credibility as a necessary condition for market efficiency. As
illustrated in chapter two, a related and underemphasized condition necessary for market
efficiency is commitment to property rights (Dales 1967).
1d) Credibility. Agencies charged with program administration must not only
be committed to enforcing rules, but in maintaining them. Glazer and Rothenberg (2001)
call this “credibility”. Implementing agents must find a suitable balance between
adjusting program rules when those rules are found to hinder program goals, and
maintaining credibility that the program will continue along a predictable path enabling
participants to plan for the future. This is especially important in programs such as cap-
and-trade where targeted firms are expected to make present investments in abatement
technologies based on the expectations that the rules will not be changed in such a way as
to alter investment decisions. SCAQMD adjustment of rules governing the reporting of
transactions would be an example of a rules change which benefits all participants in the
program. However, if program rules are changed too often then participants may lose
faith in the overall credibility of the program and of the implementing agency. While
each individual change of rules may, theoretically, be a good one, the cumulative effect
of these rule changes could be negative (Glazer and Rothenberg 2001).
Testing the credibility of the SCAQMD in the RECLAIM program is problematic.
To really get a sense of how regulated firms have reacted to the many rules changes
introduced by the SCAQMD one really would need to conduct a survey of decision
makers who have a stake in the program. Credibility, after all, is about perception.
However, a few inferences can be made about RECLAIM’s credibility in the eyes of

150
targeted firms from publicly available documents and by the behavior of those firms.
The data presented in chapter four shows that the behavior of RECLAIM facilities and
non-RECLAIM facilities did not significantly differ over the period of time examined.
One possible explanation of this is that the SCAQMD has a long history of allowing
firms that pollute more than expected to offset emissions levels through mitigation
programs. This was true before the RECLAIM program was initiated and continues to be
true today. For instance, during the 1980s large-scale polluters were allowed to continue
polluting at the same levels by offsetting that pollution through scrapping old vehicles.
When best available technology fixes proved difficult to find for regulators, they instead
allowed facilities to reduce pollution elsewhere while continuing to emit at current levels.
As shown in chapter four, throughout RECLAIM’s life cycle similar mitigation
programs have produced additional RTCs. This, along with high initial RTC allocations
meant that firms probably did not believe the RECLAIM “cap” on emissions was firm.
Their experience prior to RECLAIM implementation had been that the SCAQMD had
allowed emissions to be offset. In the early years of the RECLAIM program the high
level of RTCs also meant that little thought had to be put into planning for future
emissions reductions. RTCs were cheap, so facilities made little or no investment into
abatement technology. An SCAQMD (2007) assessment of NOx facilities found that few
investments into abatement technologies were made prior to 2000. If and when RTCs
ever were in short supply, experience had shown that the SCAQMD was willing to work
with firms in finding alternative methods for abatement.

151
The language found in SCAQMD Rule 2015, often called the “backstop
provision”, illustrates why RECLAIM’s emissions cap probably never seemed credible in
the eyes of market participants. This rule calls for a reevaluation of the program should
the price of an RTC ever hit the threshold level of $15,000. Should the price of an RTC
ever get too high, the SCAQMD was expected to make adjustments to the program to
bring those prices down. The backstop trading price was hit in late 1999, and in 2000 the
average price of NOx per tone was $45,609. Indeed, it was this backstop provision which
led to the bifurcation in the market between power-producers and non-power producers.
Other rules changes, such as retrofitting the engines on short haul ships in the coastal
waters adjacent to Los Angeles, increased the number of RTCs generated outside the
RECLAIM bubble in ways very similar to what the SCAQMD had done prior to the
adoption of RECLAIM (see SCAQMD Rule 2008). In addition, rules were changed
allowing some RECLAIM facilities to generate “RTCs” which could not be traded on the
open market, but which could offset facility specific emissions. These non-tradable
RTCs relieved these facilities of the need to purchase RTCs on the open market, lowering
demand.
While the specifics of the rules changes could not be predicted, the backstop
provisions of Rule 2015 signaled to targeted facilities that the “cap” in the program was
never credible. In fact, this is one of the general criticisms leveled at the SCAQMD by
the EPA (2001). The SCAQMD (2007, III-1-2) responded by noting that these additional
credits provide a safety valve for reducing prices and that any facility participating in
these offsets must promise future reductions in emissions. While the SCAQMD may

152
have reasons for not sticking to the cap, this amounts to an admission that they never
were committed to enforcing what is perhaps the key component of the program. RTC
prices are the driving force behind a firm’s decision to invest in abatement technology
which would lead to a lessening of emissions. It also suggests that program participants
had every reason to believe that the cap was not credible.
1e) Commitment to property rights. In a cap-and-trade market the tradable
permit is the right to emit a given amount of pollution. A tradable permit, in theory,
meets all of Dale’s (1967) conditions of property. The tradable permit in RECLAIM, the
RTC, does not. In order for a cap-and-trade market to work efficiently, there must be
faith that the property right vested in the tradable permit is secure. Without this market
participants will be reluctant to buy and sell since they have little reason to believe that
the value of the tradable permit will be maintained. Is the SCAQMD committed to
property rights? The behavior of the organization suggests that the answer is no.
Allowing facilities to offset emissions outside the RECLAIM universe, either through
mitigation programs or through the generation of additional RTCS, dilutes the value of
each RTC held. While this does have the affect of lowering overall prices, it also makes
it difficult to assess the value of an RTC given that no known cap on their production
exists. This uncertainty undermines the credibility of the program.
But we need not simply infer from the SCAQMD’s behavior that they are
uncommitted to maintaining property rights, they make the claim themselves (emphasis
mine):

153
To preserve flexibility to amend the program, suspend or terminate credits, or
even abolish the program if it is not working, it is necessary to design the credit
not to constitute a property right. This is done by including appropriate rule
language in the credit definition, and reserving the right to amend or terminate
credits or the program (SCAQMD 2007, III-1-2).

The SCAQMD explicitly state that no property right exists in an RTC. This is a key
element in maintaining an efficient market, especially when that market relies solely
upon government in its creation and in its maintenance. As outlined in chapter two, this
lack of commitment to property rights translates into diluting the value of an RTC and
should lead to an underinvestment in pollution abatement. As the data in chapter four
shows, those under RECLAIM’s cap did not significantly alter their emissions levels
differently than those in the command-and-control cohort. One explanation for why this
happened is that RECLAIM’s cap was never seen as credible and that a lack of
commitment to property rights has led to undervalued RTCs and an inefficient market.
As long as the SCAQMD is reluctant to view an RTC as kind of property, the RECLAIM
program, as implemented, can never produce the expected results predicted in the
economic literature.
Even when all the necessary ingredients for an efficient market are present, the
way in which that market is implemented might lead to inefficiencies. Is the RECLAIM
market efficient? It has a relatively large number of market participants, although that
number is significantly lower than a cursory glance might suggest. Transaction costs
seem to be low and the SCAQMD has worked to lower them even further. Enforcement
and monitoring problems have not been widespread. Yet the market may be much less

154
efficient than previous studies have concluded given that the SCAQMD lacks a
commitment to maintaining the rules and to property rights. Instead, the SCAQMD is
committed to maintaining low RTC prices, regardless of if that commitment leads to a
more inefficient market and thwarts the mechanism which theoretically drives the market
to produce desired outcomes---specifically lower emissions at a lower cost and which
inspires technological changes.
Given the choice of maintaining the market’s integrity and dealing with politically
powerful interests who wished to keep mitigation costs to a minimum, the SCAQMD
seems to have chosen the latter. As Thompson (2000) notes, the initial big supporters of
the RECLAIM NOx and SOx markets were energy producers and oil refineries. The
local natural gas monopoly opposed the adoption of the cap-and-trade program.
Thompson claims that this opposition from the natural gas company was in their
economic self-interest and, ironically, was primarily motivated by the belief that the
program would produce efficiency gains. These efficiency gains would be relatively
greater for electricity producers. This would lead to lower electricity prices causing
consumers to switch from natural gas to electricity. As the data presented in this chapter
suggest, these ex ante predictions of efficiency gains under RECLAIM are probably
overstated.
These three big players defined the politics behind the program’s adoption and
continue to be the major players in the market as implemented. Who benefited from the
bifurcation of power producers from the rest of the market? Power producers, since they
would then be allowed to “create” RTCs through creative offsets which would not need

155
to be bought on the open market. This lowered their overall costs since they would not
be paying the market price for pollution abatement which, during the power crisis, was
significant. It also helped the various oil refineries found in and around the ports of Long
Beach and Los Angeles. These oil refineries were net buyers of RTCs. As power
companies bid up the price of RTCs, oil companies and other buyers were forced to pay
higher and higher prices.
Who was hurt by this bifurcation? All RECLAIM facilities that were sellers in the
marketplace. By removing power plants from the market, demand was lowered and the
value of RTCs was reduced. Any firm that had invested in pollution abatement
technology saw a dilution of their investment. And if few firms found the cap credible
from the outset, then it seems unlikely that many of them would be motivated to invest in
the first place. Indeed, little evidence of investment into abatement technologies prior to
the rise in RTC prices in 2000 was found by the SCAQMD (2007).
The same principle applies to all of the various mitigation schemes created by the
SCAQMD to allow facilities to escape the RTC cap. More RTCs means the value of
each RTC is diluted. Mitigation programs mean that firms that would be forced to buy
RTCs on the open market, no longer demand RTCs. The price drops for other firms who
still need to buy RTCs, but the prime motivation for investing in pollution abatement
equipment is removed. As long as the two major politically powerful players—power
companies and oil refineries— are net RTC buyers, then it seems unlikely that the
SCAQMD will ever become committed to property rights and to the rules.

156
If Thompson (2000) is correct, then one of the major reasons for adopting the
RECLAIM program was that those who supported it were well organized and politically
powerful. When the rules of the program no longer benefitted them, those rules were
changed. It remains to be seen whether the less politically powerful smaller firms that are
RTC sellers can ever be organized in such a way as to influence SCAQMD to protect
their interests. The evidence suggests not. But it could be that many of these firms look
favorably on the SCAQMD’s willingness to allow mitigation in lieu of a firm cap since
the firm that is today the seller of RTCs could be tomorrow’s buyer of RTCs. An
alternative explanation might be that the SCAQMD’s credibility is so low in the eyes of
smaller firms holding RTCs that they have simply opted out of participating in the
political process which shapes those rules. Thompson (2001) found that consumer
groups were not represented by a political entrepreneur in the crafting of the initial
RECLAIM rules, but that retailers effectively blocked provisions in the rules which
would have adversely affected the welfare of consumers. One possible avenue for future
researchers would be to carefully analyze who participates in RECLAIM stakeholders
meetings and where political contributions to the local elected politicians who constitute
the SCAQMD’s Board of Directors come from.

2) Administrative efficiency. Administrative costs are incurred by regulatory agencies for
emissions measurement, monitoring, and regulating permit transactions (OECD 1997).
They are the overall direct cost to the taxpayer for the program. Calculating the cost of
the administration of the RECLAIM program is difficult. Many staff tasked with

157
overseeing the program are also tasked with overseeing other programs. RECLAIM’s
administrative budget is therefore entangled with other programs administered by the
SCAQMD. A robust analysis is therefore not offered in this study, but some preliminary
inferences can be made from the data presented in chapter four and in this chapter’s
discussion about market efficiency.
In theory a cap-and-trade market should be cheaper to run than a command-and-
control system for a number of reasons. First, cap-and-trade markets tend to be simpler
to administer since it is up to the targeted firm to come up with abatements strategies.
Under command-and-control, regulators are tasked with finding best available
technology. The more heterogeneous regulatory targets are the more difficult and costly
this task becomes. Thompson (2001, 654) notes that one of the main reasons the
SCAQMD initially supported the RECLAIM program was that they believed it would be
cheaper to administer. In fact, RECLAIM as implemented has probably been much more
expensive than cap-and-trade in theory. This is because instead of enforcing the cap and
letting prices rise to a level which would force firms to invest in technology, the
SCAQMD has opted to resorting to rule change after rule change, exception after
exception, and mitigation program after mitigation program. Each of these rules changes
has an administrative cost associated with it as staff are tasked with coming up with
recommendations and reporting on their feasibility to policymakers on the SCAQMD’s
Board of Directors.
In addition, facilities taking new mitigation options are required to draw up
detailed plans showing how they will reduce emissions in the future (SCAQMD 2002).

158
SCAQMD staff are then required to analyze whether these plans will work and if
facilities are using best available technology. The benefit of these mitigation and RTC
generation programs, as seen in the previous section, is that RTC prices are reduced for
buyers in the market. However, one of the costs is that the cap-and-trade program begins
to look more and more like the command-and-control program it replaced.
Administrative costs under RECLAIM begin to approach the administrative costs found
under command-and-control as SCAQMD staff members are asked to insert themselves
more and more into the operations of affected facilities.
The extent to which RECLAIM is administratively efficient is not known. The
data presented in this and the previous chapter suggest that it is probably much less
efficient than it would be had the cap been strictly enforced. The more RECLAIM
resembles command-and-control, the more it will cost to administer.

3. Compliance efficiency. Compliance efficiency is, “the additional cost incurred to
comply with some environmental or other legal requirement” (Ellerman et. al. 2000:
223). Compliance costs include administrative and managerial costs to the firm in
meeting regulatory standards. As noted in chapter three of this work, when regulatory
standards are process outputs, compliance costs may be extremely high and produce few
beneficial terminal outcomes. One justification for cap-and-trade is that firms are in a
better position to find the least costly way to reduce emissions than regulators tasked with
finding and forcing best available technology.

159
The extent to which RECLAIM is compliance is efficient is difficult to ascertain.
Prior to the 2000-2001 power crisis, SCAQMD staff routinely claimed in yearly
assessments that low RTC prices were an indication that firms were finding low cost
ways of complying with caps on emissions. However, as the data presented in chapter
four shows, the initially high allocation of RTCs meant that little abatement actually took
place. The rules changes following the power crisis also make this task more difficult.
Under the ideal of cap-and-trade, the clearing price of an RTC would be equal to
marginal abatement costs. Because RTCs would be sold across firms, compliance costs
would be equalized throughout the program. But because new rules allow some firms to
forego buying RTCs in the open market and instead mitigate emissions, it is difficult to
tell just how much it costs for firms to comply.

Innovation

Has RECLAIM produced innovation in pollution abatement? No.

The last major improvement that economists predict for cap-and-trade markets is
that they lead to innovation that would not have been found under command-and-control.
This is based on the theory that firms will actively seek the least costly way to comply
with strict caps on the amount of emissions allowed. Firms seeking to lower costs will
find technology or process changes that lower emissions. If the price of tradable permits
is high enough, an entrepreneurial firm with low marginal abatement costs can make a

160
profit by selling extra permits on the market. The prediction is that cap-and-trade
markets are better suited to forcing innovation than are command-and-control systems
which rely on centralized actors finding best available technology.
The causal mechanism by which this innovation is achieved is price. The higher
the price of a tradable permit, the more motivated firms will be to find new ways of
lowering emissions. In late 1999 the price of RTCs for NOx began to dramatically
increase as supply finally met demand. The SCAQMD understood this rise in the price
of RTCs as a failure in the market. It is not clear, though, why this increase in price
should be considered a market failure. No theory exists to support this conclusion. Price
is the causal mechanism in a cap-and-trade market for lowering emissions. High prices
are not a market failure, but are an incentive to firms to find cheaper ways to lower their
emissions. In order for the market to properly signal to those with lower marginal
abatement costs to produce more RTCs, the price must be sufficiently high. Otherwise,
there is no incentive for any firm to lower their total emissions.
Data in chapter four show that prior to 2000 that the vast majority of RTCs traded
were without price. This is an indication that the most prevalent way in which firms
chose to lower emissions was to “trade” RTCs from one facility they owned to another
facility. This is not necessarily a bad thing. Again, the point of a cap-and-trade market is
to allow firms to decide how they will reduce emissions. If those reductions come from
closing a less efficient facility and running a more efficient facility at peak capacity, there
is no substantial loss to the environment nor is their a reduction efficiency gains predicted

161
by economic theory. If this is the cheapest way to abate, then this is what firms ought to
do. And in fact, this is what the data presented in chapter four suggest happened.
However, as demand for power increased and power suppliers found a need to
operate facilities with higher pollution levels at higher capacity they could no longer
simply move RTCs from one facility to another without price. Theory predicts that it is
only at this point that innovation will occur. As RTC prices were bid up, entrepreneurial
firms participating in RECLAIM would only then began to search for innovative ways to
reduce pollution. Clearly the SCAQMD is committed to keeping the price of RTCs low,
but not in allowing the market to work as intended. In a self evaluation of the program the
SCAQMD (2007 III-1-3) blamed firms on the lack of investment in innovative
technologies, claiming that, “some facility managers exhibited short-term thinking and
intentionally delayed capital expenditures for emission reduction projects in order to
increase profits in the near-term.”
It is difficult to see why facility managers are to blame if the SCAQMD truly
understands cause and effect. In the case of command-and-control what causes
investment into emission reductions projects is that firms are forced to do so. In the case
of cap-and-trade capital expenditures for emission reductions projects are made only
when the price of a permit is high enough to justify that investment. The SCAQMD
holds the untenable position that high RTC prices are a problem, when in fact they are the
solution. Without sufficiently high RTC prices firms simply are not motivated to
innovate.

162
It is not surprising, therefore, that the SCAQMD (2007) finds little evidence of
innovation among facilities regulated under the RECLAIM program. In response to the
EPA (2007, 23) finding that, “Most facilities did not … make new capital expenditures,”
the SCAQMD (2007) claimed that at least some facilities have updated equipment or
implemented process changes, but also admit that “it is difficult to fully quantify this
aspect of the program” (III-1-4). Initial allocations were problematic, which may have
contributed to underinvestment in abatement technologies, but they place the blame on
the shoulders facility managers:
a sense of complacency by many facilities [that] reduced the pressure to install
controls. Models of the program assumed rational economic behavior, where
facility owners and operators would add controls when it was to their economic
advantage, but this did not always occur (EX II).

What is surprising about this conclusion about why few facilities made efforts to
install new emissions control equipment is that they seem not to understand that it was
poor program design and implementation that caused this, not irrational behavior. In fact,
the choice to forego technological improvements is extremely rational when RTC prices
are low. As discussed in the previous chapter, far too many RTCs were initially granted
in the first years of the program’s life. This was due to poor policy design. However, as
Thompson (2000) shows, the grandfathering in of emissions levels at peak levels may
have been necessary in order for the SCAQMD to garner enough political support to see
that the policy was adopted in the first place. When prices began to go up in 1999, if one
of the program’s goals was to inspire innovation, then the proper response would have
been to allow the price mechanism to work. There is no reason to believe that firms

163
would invest in innovative technology unless the price of RTCs was well above average
prices from 1994-1999.

Conclusion

This chapter has focused on the extent to which the RECLAIM program has met
expectations predicted by economic theory. RECLAIM seemed to be a good candidate
for a successful cap-and-trade program because the market seems, at first glance, to be
relatively large and its targets are relatively heterogeneous. However, a closer look
reveals that RECLAIM is not nearly as large as previously suggested. RECLAIM has
met statutorily outlined goals, but it has not led to emissions reductions beyond what
were found in command-and-control. The extent to which RECLAIM has been more
efficient than the command-and-control alternative is not clear, but it is likely that
RECLAIM is not as efficient as economic theory predicts it could be. The last goal of
cap-and-trade programs is to produce innovation, and clearly RECLAIM has failed to do
this.
As discussed in this and the previous chapter, many of RECLAIM’s shortcomings
can be attributed to initial high RTC allocations and various rules changes introduced
over the life of the program. These rule changes signal that the program caps on
emissions were never credible and that the SCAQMD is not committed to property rights.
The majority of these rules changes were meant to manipulate the price of RTCs
downward. To that end these rules changes have met their stated purpose. RTC prices

164
began to decline in the years following the power crisis, and today RTC supply is well
above RTC demand.
As outlined in chapter two, implementation studies seek to explain the link
between programmatic goals and programmatic achievements (Knaap and Kim 1998).
The most prevalent finding in the first generation of implementation research was that
policy outcomes are either disappointing or unwitting, largely because of the
incompatible goals or values of those assigned the task (Calista 1994). The second
generation was more optimistic about the prospects of policy implementation because a
well-structured statute can limit the ability to thwart that policies’ intention by
uncooperative agents or those targeted by the policy. “Bottom up” scholars also
emphasized the need for flexibility from those tasked with implementing a policy, but
that implementing agents must also be committed to programmatic goals (Hill and Hupe
2002).
RECLAIM, as implemented, is a far cry from the ideal cap-and-trade market
imagined in economic theory. But given that dozens of previous implementation studies
have found that policies as implemented on the ground are often quite different than the
policy imagined in the halls of Washington, the state capital, or city hall, this should
come as no surprise. These changes come as programmatic ideals are confronted with
political, economic, and even environmental realities which create the context in which
the policy goes from idea to working on a day to day basis. In the case of RECLAIM, the
ideal cap-and-trade market was thwarted from the beginning. Part of this was simply
because the Los Angeles basin offers a unique environmental context. Air Pollution in

165
the basin is the worst in the country partly because inland mountains trap emissions that
in other areas are simply blown away. This led to splitting the RTC market spatially and
temporally. The market in practice is therefore much smaller than many have presumed.
Another deviation from the ideal during the initial formulation stages was the high
level of RTCs allocated to participating firms. As Thompson (2001) notes, this was
because if the policy had any hopes of being passed in the first place, then it became
politically expedient to raise initial RTC allocations to a level where affected groups
would not be confronted with lowering emissions until several years after the program
had begun.
Even given this initial problem in programmatic design, the policy still had the
potential to live up to the promises of cap-and-trade, but given initial levels of RTCs,
these promises were just pushed to a future date. The ingredients for a cap-and-trade
program to be efficacious in meeting emissions goals, efficiency goals, and produce
innovation were all there. The RECLAIM statute initially structured incentives in such a
way that targets should have produced the desired outcomes (Mazmanian and Sabatier
1983). What thwarted some of these goals in RECLAIM are the same things that thwart
programmatic goals in more traditional programs. The conflicting goals of those tasked
with implementing RECLAIM program are at the heart of this problem. This is
consistent with early implementation studies which found that those involved in the
policy process often had incompatible goals and values (Calista 1994). The SCAQMD
wishes to both keep RTC prices low and have an efficient market which produces
effective outcomes and which produces innovation. But these goals are incompatible

166
since the causal mechanism behind innovation and therefore effectiveness is price. As
long as the SCAQMD is committed to keeping prices artificially low through various
schemes, we should not expect to see many of the benefits attributed to cap-and-trade
come to fruition.
Given the political context which the RECLAIM program found itself in, though,
perhaps these incompatible goals are inevitable. Those who directly benefit from low
prices are precisely the same groups who initially supported the program’s inception
(Thompson 2001). Two of the “big three” groups that pushed for the program’s
implementation, electrical utilities and oil refineries, benefit from rules changes reducing
the price of RTCs. As net buyers of RTCs they have an incentive to pressure the
SCAQMD to continue to ignore the hard caps on emissions levels initially planned in the
RECLAIM program and which are universally thought of as a necessary ingredient for a
cap-and-trade program. Unless those adversely affected by low RTC prices can find a
political entrepreneur to mobilize their interests, we should expect to find the SCAQMD
continuing in its support of the mitigation programs designed to keep those prices low.
One interesting finding from 12 years of RECLAIM implementation experience
conflicts with that of Thompson (2000). Thompson notes that the SCAQMD supported
RECLAIM initially because it was easier for them to administer. Recall that RECLAIM
is well suited as a critical case study because targets for regulation are relative
heterogeneous and regulators has found it difficult to identify readily available
technological fixes. The fact that staff at the SCAQMD supported the program was

167
evidence to Thompson that they realized the conditions were right for a cap-and-trade
program to make the most sense. Adopting RECLAIM would mean less work for them.
This is contrary to the predictions of some public choice theorists who claim that
regulators have an organizational incentive to support command-and-control because it
requires them to do more work and thus grow the organization (Brady et al 1990). But if
the rules changes introduced since 2000, the year Thompson’s article was published, are
any indication, then the SCAQMD has found plenty of extra work and job security. The
administration of the program is far more complicated today than it was eight years ago.
The agency is no longer simply tasked with monitoring and enforcement, but also with
new mitigation programs and even overseeing the same best available technology
provisions found in command-and-control. The agency continues to sponsor rules that
also directly benefit the growth of the organization, just as public choice theory predicts it
would. Further research into the organizational incentives and behavior at the SCAQMD
in light of the RECLAIM experience might produce some interesting results for public
choice and organizational theorists.
As the years progress, the RECLAIM program deviates farther and farther away
from its cap-and-trade ideal and takes on many of the characteristics which could be
found in command-and-control programs administered by the SCAQMD prior to its
inception. The RECLAIM program, as implemented, is different than cap-and-trade
programs as theorized. Therefore, we should expect to see different outcomes then those
theories would predict. This is exactly what the data here and in the previous chapter
shows. The SCAQMD also had a long history of allowing flexibility in compliance, as

168
seen by the various mitigation programs they sponsored. If this practice has continued,
then those in the control group used in chapter four also don’t represent command-and-
control in its pure form. There seems to be an indication that these policies are
converging towards some middle ground between a pure cap-and-trade program and a
pure command-and-control program. RECLAIM, as implemented, is not the cap-and-
trade ideal, but neither is command-and-control in the same region. The following
chapter discusses the overall implications of the findings presented in this study.

169
Chapter 6 Discussion and Conclusion

Implementation: between theory and practice

This study was undertaken in an atmosphere where cap-and-trade markets are
widely advocated. Since each increment of environmental improvement seems to
becoming more and more costly under traditional command-and-control approaches,
innovations such as cap-and-trade are widely seen as a way to break the policy impasse
and continue to improve environmental quality in a more efficient manner (Portnoy and
Stavins 2000). The diffusion of cap-and-trade markets is happening fast. Led by the
EPA, the Title IV acid rain program has served as the example by which new air quality
programs should be emulated (Ellerman et al. 2000). The European Union has adopted a
cap-and-trade program by which member countries hope to implement the Kyoto
protocol largely based on the successful Title IV program.
RECLAIM has also been cited as a model for the potential benefits of cap-and-
trade over command-and-control. The theoretic literature suggests that as regulatory
targets become more diverse that command-and-control becomes less efficient. Cap-and-
trade markets should be most successful when two criteria are met. First, when there is a
heterogeneous target population. Targets for RECLAIM regulation are far more diverse
than those in the Title IV program. Second, they are well suited when regulators find it
difficult to identify readily available technological fixes. The data presented in chapter
five of this work showed that in the initial formulation stages of RECLAIM, staff at

170
SCAQMD believed it would be difficult for them to find best available technologies for
the variety of facilities which would need to be regulated under the existing AQMP.
Thus, RECLAIM also meets the second criteria. RECLAIM should have produced the
benefits predicted.
Earlier studies (Kamieniecki, Shafie, and Silvers 1999) have shown that
RECLAIM also met the three conditions necessary for an efficient cap-and-trade market.
RECLAIM has many participants, although this number has probably been overstated in
past studies; transaction costs are low and becoming lower as the market becomes more
transparent; and there is sufficient monitoring and enforcement. If all the ingredients for
success are present in RECLAIM, then what does one make of the program’s apparent
shortcomings?
The harshest critics of the of RECLAIM program have been staff at the EPA. The
general feeling one gets from reading the EPA’s 2002 assessment of the RECLAIM
program is this: how could the SCAQMD mess up cap-and-trade so badly? It should be
remembered that many who had declared RECLAIM a “failure” took place even as the
Bush Administration’s EPA was pushing for the adoption of more cap-and-trade
programs, such as the “Blue Skies” initiative. It also came at a time when the European
Union was considering adopting cap-and-trade. Perhaps most important still,
RECLAIM’s alleged failure came precisely as some environmental group’s were pushing
cap-and-trade as a possible way to lure a wary American public into approving the Kyoto
Treaty on climate change. If lowering greenhouse gasses could be done at a much lower
cost and in a less intrusive way, perhaps more Americans would be willing to support it?

171
RECLAIM presents an obstacle for the diffusion of cap-and-trade. It has been a
thorn in EPA’s side. In fact, it has been a thorn in the side of many in the environmental
movement who also had a stake in seeing cap-and-trade diffused. Market based
instruments are alluring to enough conservatives and Republicans that pushing them has
been seen as a way to break decades long gridlock which had stalled proposed
environmental policies at all levels of government (Bryner 1995). So, how to explain the
apparent failure of a cap-and-trade market to produce some of the desired outcomes?
Blame the implementing agency. Cap-and-trade markets, per se, had not necessarily
failed because RECLAIM had been so poorly designed and implemented.
RECLAIM continues to be a flawed cap-and-trade market. Contrary to the claims
of the SCAQMD and of others who have attempted to assess the policy, it is not a cap-
and-trade market in the sense that it operates in the form as advocated by the economic
literature. But neither is it a command-and-control program. It is a hybrid of the two.
Unfortunately, the most important aspects of cap-and-trade have been abandoned during
the implementation process: strict adherence to the cap and allowing the price mechanism
to work.
As we have seen, two important conditions necessary for a successful cap-and-
trade market have been overlooked by previous studies. Those conditions are committed
to the rules and to property rights. It is likely that program participants never believed
that caps would be strictly enforced and that they would be allowed to offset facility
emissions just as they had been allowed to do before RECLAIM’s adoption. The initial
statutory language contained in Rule 2012 (the “backstop provision”) suggests that the

172
SCAQMD had always planned to revisit program rules if prices ever got too high. The
SCAQMD is committed to cap-and-trade in theory, but not in practice. To be committed
to cap-and-trade must, by definition, mean to be committed to the mechanism by which
the benefits of cap-and-trade are realized. That mechanism is price. The SCAQMD’s
actions in allowing offsets, mitigation, and outside the bubble RTC generation show that
they are not committed to property rights. This is readily admitted by them in their own
internally generated reports.
That RECLAIM in practice is a far cry from cap-and-trade in theory should come
as no surprise. Much of the implementation literature predicts just that. As policies are
translated from intent to practice they are changed. Much of this change is dependent on
the context of the policy environment. To some extent, the unique physical environment
of the Los Angeles basin makes conditions for a successful market less likely. In the case
of RECLAIM, the same powerful political actors that initially supported the program also
benefited from each of the changes introduced during implemented which made it deviate
farther and further from the cap-and-trade ideal.
However, for all the flaws in the policy pointed out in chapter five, the data
presented in chapter four are a stark reminder that if predictions about the likely path of
RECLAIM were overly optimistic, so too were predictions about the command-and-
control alternative. In the same years that RECLAIM facilities were polluting, on
aggregate, above and beyond what was legally allowed, command-and-control facilities
were emitting above and beyond what was predicted. This study shows that for whatever
flaws might be found in cap-and-trade markets – even ones that had apparently “failed”

173
to meet certain expectations – flaws may also be found in the next most likely alternative,
command-and-control. If RECLAIM had failed to be as effective as predicted, so too
had command-and-control.

Designing better instruments for measuring policy effectiveness

The most important contribution of this study is that it offers new ways in which
to measure the relative impact of cap-and-trade. What all of those who have criticized
the effectiveness of RECLAIM have taken for granted was that the way we were
measuring success was based upon a flawed assumption about the path command-and-
control would have taken. As the data in chapter four make clear, the SCAQMD
counterfactual was wildly pessimistic about how firms would have behaved under
command-and-control in the early years of the program, and then wildly optimistic about
how command-and-control would have effected emission levels in later years. For
instance, in 1995 the SCAQMD reported that 28% of NOx RTCs went unused (see Table
4.1). Apparently RECLAIM had worked much better than command-and-control. How
much better? 28%. But what was the actual difference between NOx emissions in that
year between those facilities operating under RECLAIM and those operating under
command-and-control? In that year, both RECLAIM and non-RECLAIM facilities
emitted exactly the same amount as they had emitted the year before. Actual difference?
0% (see Table 4.6).

174
So, perhaps RECLAIM was a failure from the beginning? Even in those years in
which earlier studies, relying on the SCAQMD counterfactual, had proclaimed it a
success? But let us look at the years in which RECLAIM was considered a “failure” by
nearly everyone who took an interest. In 2000 RECLAIM was alleged to have emitted
19% more NOx than it would have under command-and-control (see Table 4.1). In that
year RECLAIM facilities were emitting 16.5% less than they had in 1993, while non-
RECLAIM facilities were emitting 20% less. While this may seem like a dramatic
difference showing that RECLAIM was not performing as well as the command-and-
control alternative, it should be remembered that the difference between the two groups
was not statistically significant at the .05 level (see Table 4.6). The very next year, 2001,
when by the counterfactual measure RECLAIM was emitting slightly more than the
command-and-control alternative (Table 4.1), the RECLAIM group was actually emitting
less. A 36% drop from the baseline year for RECLAIM with only a 24% drop from the
baseline year for the non-RECLAIM group (Table 4.6).
Again, these differences were not statistically significant, but from this lack of
difference we can infer that whatever was causing falling levels of emissions for facilities
in the RECLAIM universe was also causing falling levels of emissions in the non-
RECLAIM universe. In other words, the cap-and-trade program – the intervention or
“treatment” – had no measurable effect on the dependent variable. One might conclude
from this that RECLAIM had indeed “failed”. But by what standard does one measure
failure or success in policy implementation? As we have seen in chapter four, the

175
baselines used for comparison matter very much in defining and operationalizing the
term.
The first goal of cap-and-trade programs such as RECLAIM is to deliver
equivalent amounts of pollution reductions as command-and-control, in a much more
efficient matter. The RECLAIM case delivers similar amounts of pollution reduction as
similarly situated facilities still operating under command-and-control. If RECLAIM has
underwhelmed many with high expectations of how cap-and-trade should perform, then
implicit in this is the notion that command-and-control also should have performed much
better. For if the baseline is command-and-control, then how can one argue that
RECLAIM has failed when the relative impacts of the program are nearly
indistinguishable from that baseline?
There is a deeper implication to this finding. Comparing the SCAQMD
RECLAIM counterfactual to how command-and-control operates shows that the
counterfactual is wrong. This means that internal estimates of program performance are
also flawed. Since the vast majority of policy researchers also use the same measure then
it stands to reason that at least some of their findings are also flawed. And if the number
of RTCs issued is not a valid metric, then what about other command-and-control
programs which are evaluated based on similar metrics?
Since tradable permits used as a command-and-control counterfactual may not be
a reliable metric, those wishing to assess the performance of market based instruments
such as cap-and-trade should find alternative methods of doing so. This study has
proposed that policy researchers should find similarly situated firms regulated under

176
alternative regimes and then compare the performance of the two groups. RECLAIM
was readily amenable to this kind of analysis since all the firms regulated under it had
also been regulated under command-and-control at some time in the past. RECLAIM
facilities also operated side by side with facilities which were not subject to the cap-and-
trade program. Making comparisons, then, was easy.
The Title IV program may not be as amenable to this type of analysis. Much of
Title IV’s success might be attributable to a very different hypothetical counterfactual.
Title IV represented a large breakthrough in air pollution regulation because it broke
through a 20 year political impasse over what to do with heavily polluting coal burning
power plants (Bryner 1995). Unlike those participating in the RECLAIM program, many
of the facilities included in the first phase of Title IV were not regulated under command-
and-control. That is, they had been specifically exempted from command-and-control
regulations under the first Clean Air Act. As Ackerman and Hassler (1981) point out,
this created perverse incentives for power companies to pollute even more by shifting
production to older and less efficient facilities which had been grandfathered out of the
CAA.
Those facilities included in the first phase of Title IV were never in the universe
of command-and-control. There is no baseline by which to measure the relative impacts
of the program. But there is another methodological implication to this implicit in
Bryner’s (1995) description of how the CAA was amended in 1990 to target these
facilities for regulation under cap-and-trade. If the Title IV program had not been
adopted then there is a good chance that these facilities would have continued to operate

177
as they had prior to 1990. That is to say, they would have remained exempt from
command-and-control regulations and therefore would have continued to be major
polluters. This suggests that the true counterfactual baseline for Title IV would not be
command-and-control, but no regulation whatsoever.
Any lessening of emissions under Title IV – no matter how small – would be an
improvement. The point of Ackerman and Hassler (1981) is that the irony of the Clean
Air Act was that in some areas the air actually became dirtier because of it. The
incentives were so perverse that companies that would normally invest in efficiency
improvements under a free and unregulated market did not. Installing newer
technologies or upgrading facilities triggered new source performance standards, nullified
grandfather clauses, and brought the facility into the regulated universe. A good case can
be made that the counterfactual for Title IV is the old CAA system of perverse incentives.
The Title IV program is universally seen as successful because the alternative was
so bad. This suggests that cap-and-trade programs might be especially appropriate when
the alternative to adopting the program is political impasse marked by no change in the
status quo. Even if cap-and-trade markets produce only marginal environmental
improvements, this may be better than a situation where no improvements are made
whatsoever. What this does not speak to, though, is the feasibility of the adoption of cap-
and-trade under such circumstances. As Thompson (2000) notes, where no prior
regulation exists it may be politically difficult to find support for cap-and-trade since the
costs and benefits of the program may not be Pareto optimal. What is overlooked by

178
Thompson is that those same equity concerns about the move from no regulation to cap-
and-trade are also present when moving from no regulation to command-and-control.
But what about the vast majority of cases where the alternative is more
ambiguous? Such is the case of RECLAIM. Firms participating in the program had
previously been regulated under command-and-control and by all accounts had already
significantly reduced emissions. Since this is more likely representative of most firms
operating within the United States, then researchers should have ample resources for
comparing how cap-and-trade markets do compared to the command-and-control
alternative. The vast majority of firms regulated for air pollution in the U.S. operate
under some form of command-and-control. Only a few of these facilities operate under
cap-and-trade. The trick is to find similarly situated facilities regulated under both
systems. I doubt that this is too tall an order for the policy studies community.
As cap-and-trade markets diffuse on a worldwide scale, so too should the
opportunities to compare and contrast them to the next regulatory alternative. Especially
interesting is the European context since other market instruments and implementation
strategies are more widespread. In the United States the general alternative to cap-and-
trade continues to be command-and-control. In Europe there is much more widespread
use of other alternatives such as effluent taxes. In addition, the regulatory environment of
Europe provides a different context to explore the ways in which the theory of cap-and-
trade is translated into practice. As we have seen in the case of RECLAIM, context
matters. The question remains whether or not the European bureaucratic model, which
stresses apolitical professionalism, adds more favorable conditions for cap-and-trade

179
markets to be successful. Other aspects of the European model, which stress flexibility,
might mean that implementing agents will lack commitment to the rules and to property
rights and will be amenable to providing outlets from the “cap” such as mitigation
programs.

Conclusion and implications

RECLAIM is just as successful at reducing the targeted air pollutants as the
command-and-control alternative, but this does not mean that those implementing the
program got it all right. The data presented in chapters four and five might be interpreted
in a number of ways. The glass half full interpretation of these data might be that cap-
and-trade is at least as successful as command-and-control. But the glass half empty
view might bet that cap-and-trade might sometimes be just as bad as the regulatory
alternative. Glazer and Rothenberg (2001) note that implementation of market
instruments are susceptible to the same types of human foibles as are any other
government endeavor. All of this leads to one important inference: market mechanisms
for regulation are just as susceptible to political interference as are other forms of
regulation. It is not surprising that politics happens at all points during the policy
process. This is the general crux of criticism of the “stages” heuristic for dividing the
policy process up into discrete units of observation. That politics happens at all stages of
the process is well known and one would be hard pressed to find any academic work that

180
would dispute it. The bifurcation of politics and public administration as distinct areas of
study also makes no sense. This is clearly not the case.
Political interference which may corrupt the effectiveness of environmental
regulations will be found in all implementation strategies, regardless of what policy tool
is chosen. What is different is the mechanism by which that interference comes. In cap-
and-trade markets scholars have long been aware that we should be particularly careful in
the initial distribution of tradable permits because initial allocations might violate
generally held norms of equity (Smith 2004). In addition, if the cap in cap-and-trade is
too high, prices will be too low, and the causal mechanism behind pollution abatement
will be thwarted (Portnoy and Stavins 2000). But in addition to these concerns, policy
scholars who are interested in alternatives to command-and-control should be keenly
aware of how political influence can be used to undermine commitment to property rights
and to the rules. A lack of credibility and to letting the price mechanism work is a recipe
for a cap-and-trade market which will not be as effective as possible.
As cap-and-trade markets spread rapidly, policy makers and policy implementers
should be keenly aware of carefully designing these programs for maximum success.
When prices for tradable permits in a cap-and-trade market rise, this does not necessarily
mean that the program is not efficient. Prices must be allowed to rise to a high enough
level and for a long enough time that they will serve as a signal for firms to make the
proper amount of investment required to abate pollution to expected levels.
For researchers this study has serious methodological implications. Measuring
how well cap-and-trade markets do is dependent upon the baseline by which one makes

181
the comparison. It is not enough to simply accept that the “cap” in cap-and-trade
sufficiently represents a counterfactual case where the program was not implemented.
There is importance in using this as a measure of success, but not the only one. Using the
internal standards of implementing agencies removes observer bias, but it introduces
agency bias and the biases built into the baseline measure introduced through the political
process which created it. There may be times, like in the Title IV program, that this is the
only baseline available since facilities in the program were not subject to prior command-
and-control regulation. But most firms in the U.S. which will be included in cap-and-
trade programs in the future will likely to have also been regulated by command-and-
control in the past. Further, these facilities under cap-and-trade will also likely be
surrounded by other facilities regulated by traditional command-and-control.
Researchers would do well to identify an empirically based control group from which
meaningful comparisons can be based.
Doing so will lead usable knowledge and will help improve programmatic
performance in a much more reliable way than in relying on counterfactuals which may
or may not have any basis in reality. If the goal of environmental policymaking is to
make actual improvements in environmental quality, then empirically based studies are
the best hope for producing usable knowledge. In this way policies can be improved,
environmental quality enhanced, and human health better safeguarded.

182
References

Anderson, R. 1997. The U.S. Experience with Economic Incentives in Environmental
Pollution Control Policy. Washington, DC: Environmental Law Institute.

Anderson, T. L. and D. R. Leal. 2001. Free Market Environmentalism. New York:
Palgrave.

Ackerman, B. A. and W.T Hassler. 1981. Clean Coal and Dirty Air. New Haven, CT:
Yale University Press.

Bardach, E. 1977. The Implementation Game: What Happens After a Bill Becomes a
Law. Cambridge, MA: MIT Press.

Barrett, S. 2000. “Political Economy of the Kyoto Protocol.” In D. Helm (Ed.),
Environmental Policy: Objectives, Instruments, and Implementation. Oxford:
Oxford University Press.

Bartlett, R.V. 1994. “Evaluating Evironmental Policy Success and Failure.” In Norman
J. Vig and M. E. Kraft (Eds.), Environmental Policy in the 1990’s. Washington,
DC: CQ Press, pp. 167-168.

Barrett, S.M. and C. Fudge. 1981. “Reconstructing the field of analysis.” In S.M. Barrett
and C. Fudge (Eds.), Policy and Action: Essays on the Implementation of Public
Policy. London: Methuen, pp. 249-78.

Bobrow, D.B., and J.S. Dryzek. 1987. Policy Analysis by Design. Pittsburgh:
University of Pittsburgh Press.

Braden, J. B. and C.G. Kim. 1998. “Economic Approaches to Evaluating Environmental
Programs.” In G. J. Knaap and T. J. Kim (Eds.), Environmental Program
Evaluation: A Primer. Urbana, IL: University of Illinois Press, pp. 203-237.
Brady, G.L., M.T. Maloney, and A.F. Abbott. 1990. “Political Limits of the Market for
‘BAT’ Medallions.” Cato Review of Business and Government Regulation 13:
61-64.

Bryner, G. C. 1995. Blue Skies Green Politics: The Clean Air Act of 1990 and Its
Implementation, 2
nd
ed. Washington, DC: CQ Press.

183
--------. 1997. “Market Incentives in Air Pollution Control.” In S. Kamieniecki, G. A.
Gonzalez, and R. O. Vos (Eds.), Flashpoints in Environmental Policymaking:
Controversies in Achieving Sustainability. Albany: SUNY Press, pp. 85 – 107.

--------. 1999. New Tools for Improving Government Regulation: An Assessment of
Emissions Trading and Other Market-Based Regulatory Tools. Arlington, VA:
PricewaterhouseCoopers Endowment for the Business of Government.

Bullock, C.S. 1981. “Implementation of Equal Opportunity Programs.” In D.A.
Mazmanian and P.A. Sabatier (Eds.), Effective Policy Implementation.
Lexington, MA: Heath.

Calista, D. 1994. “Policy Implementation.” In Encyclopedia of Policy Studies, S. Nagle
(Ed.). New York: Marcel Dekker, pp. 117-155.

Cobb, R. W., and C. D. Elder. 1983. Participation in American Politics: The Dynamics
of Agenda-Building, 2
nd
ed. Baltimore: The Johns Hopkins University Press.

Colby, B.G. 2000. “Cap-and-Trade Policy Challenges: A Tale of Three Markets.” Land
Economics 76: 638-58.

Coase, R.H. 1960. “The Problem of Social Cost.” Journal of Law and Economics 3:1-
44.
Coglianese, C. and L. D. Snyder Bennear. 2005. “Program Evaluation of Environmental
Policies: Toward Evidence-Based Decision Making.” In G. D. Brewer and P. C.
Stern (Eds.), Decision Making for the Environment: Social and Behavioral
Science Research Priorities. Washington, D.C.: The National Academic Press
(Appendix E, pp. 246-273).

Dales, J.H. 1967. “The Property Interface” Reprinted in R. Dorfman and
N. S. Dorfman (Eds.), Economics of the Environment, 2
nd
ed.
(1977), pp. 172-186.

Darst, R. G. 2001. Smokestack Diplomacy: Cooperation and Conflict in East-West
Environmental Politics. Cambridge: M.I.T. Press.

deLeon, P. 1998. “The Missing Link Revisited: Contemporary Implementation
Research.” Policy Studies Review 16: 311-338.

184
---------. 2001. “A Democratic Approach to Policy Implementation.” Paper presented at
the Annual Meeting of the American Political Science Association, San Francisco,
CA, September 2001.

deLeon, P. and K. Kaufmanis. 2000. “Public Policy Theory: Will it Play in Peoria?”
Policy Currents 10: 4 (Winter), pp. 9-12.

deLeon, P. and C. R. Martel. 2003. “Policy Sciences Approach.” In J. Rabin (Ed.), The
Encyclopedia of Public Administration and Public Policy, Volume II. New York:
Marcel Dekker, pp. 939 – 942.

Derthick, M. 1972. New Towns, In-Town. Washington, D.C.: Urban Institute.

Dobra, J.L. and J. Wendel. 1999. “Pursuing Environmental Goals: Regulations Versus
Markets.” In D.L. Soden and B.S. Steel (Eds.), Handbook of Global
Environmental Policy and Administration. New York: Marcel Dekker, pp. 81-
111.

Eckstein, H. 1975. “Case study and Theory in Political Science.” In F. Greenstein and
N.W. Polsby (Eds.), Handbook of Political Science. Reading, MA: Addison-
Wesley, 52-53.

Ellerman, A.D, R. Schmalensee, E.M. Bailey, P. L. Joskow, and J.P. Montero. 2000.
Markets for Clean Air: The US Acid Rain Program. New York: Cambridge
University Press.

Environmental Protection Agency (EPA). 2002. An Evaluation of the South Coast Air
Quality Management District’s Regional Clean Air Incentives Market – Lessons
in Environmental Markets and Innovation. Washington, D.C.: Author.

Ferman, B. 1990. “When Failure is Success: Implementation and Madisonian
Government.” In D.J. Palumbo and D. Calista (Eds.). Implementation and the
Policy Process: Opening Up the Black Box. Westport, CT: Greenwood Press.

Foreman, C. 1998. The Promise and Peril of Environmental Justice. Washington, DC:
Brookings.

Freeman, A. M., III. 2003. “Economics, Incentives, and Environmental Policy.” In
Norman Vig and Michael Kraft (Eds). Environmental Policy, 5
th
ed.
Washington, DC: CQ Press.

185
Gangadharan , L. 2001. “Compliance in Environmental Markets.” Applied Economics
Letters 8: 641-644.

Gerring, J. 2004. “What is a Case Study and What is it Good for?” American Political
Science Review 98: 341-354.

Glazer, A. and L.S. Rothenberg. 2001. Why Government Succeeds and Why It Fails.
Cambridge: Harvard University Press.

Goklany, I.M. 1999. Clearing the Air: The Real Story of the War on Air Pollution.
Washington, DC: CATO Institute.

Goggin, M.L., A. O’M. Bowman, J.P. Lester, and L.J. O’Toole, Jr. 1990.
Implementation Theory and Practice Towards a Third Generation. Glenview, IL:
Scott, Foresman/Little, Brown.

Gormley, W.T., Jr. 2000. “Environmental Performance Performance Measures In a
Federal System.” Learning from Innovation in Environmental Management,
Research Paper No 13. National Academy of Public Administration: Washington,
DC.

Gormley, W.T. Jr. and D. Weimer. 1999. Organizational Report Card. Cambridge:
Harvard University Press.

Greenberg, R. 1997. We Can Save the World: The Uniworld Plan. Van Nuys: Metro-
West Publishing.

Hahn, R.W., and R.G. Noll. 1982. “Designing a Market for Tradable Emissions
Permits.” In W.A. Magat (Ed.), Reform of Environmental Regulation.
Cambridge, MA: Ballinger Press, pp. 119-146.

Hahn, R.W. and J. Hird. 1991. “The Costs and Benefits of Regulation: Review and
Synthesis.” Yale Journal of Regulation 8: 233.

Hall, J.V., and A.L. Walton. 1996. “A Case study in pollution markets: Dismal science
versus dismal reality.” Contemporary Economic Policy 14: 67 – 78.

Hamilton, C. R., and D.T. Wells. 1986. The policy puzzle. Englewood Cliffs: Prentice
Hall.

Harrison, D. and A. Nichols. 1990. Market-Based Approaches to Reduce the Cost of
Clean Air in California’s South Coast Basin. Diamond Bar, CA: SCAQMD (self
published report).

186

Hargrove, E.C. 1975. The Missing Link: Implementation in Social Policy. Washington,
DC: The Urban Institute.

Hawkins, J., S. Lowe, G. Simon, and N. Suetake. 2001. An Evaluation of the Los
Angeles Regional Clean Air Incentives Market. Master’s thesis, University of
California, Santa Barbara.

Helm, D. (Ed.). 2000. Environmental Policy: Objectives, Instruments, and
Implementation. Oxford: Oxford University Press.

Hill, M. and P. Hupe. 2002. Implementing Public Policy. London: Sage.

Hix, M., E. Ruder, and D. Sugarman. 2000. “Analysis of Volatile Trading Systems Air
Pollution Organic Compound.” Learning from Innovations in Environmental
Protection Research Paper 5. Washington, D.C.: National Academy of Public
Administration.

Hjern, B and C. Hull. 1983. “Implementation Research as Empirical Constitutionalism.”
European Journal of Political Research 10: 105-115.

Ingersoll, T.G., and B.R. Brockbank. 1986. “The Role of Economic Incentives in
Environmental Policymaking.” In S. Kamieniecki, R. O’Brien, and M. Clarke
(Eds.), Controversies in Environmental Policy. Albany: SUNY Press, pp. 201-
222.

Ingram, H. 1990. “Implementation, A Review and Suggested Framework.” In N. Lynn
and A. Wildavsky (Eds.), Public Administration: The State of the Art. Chatham,
NJ: Chatham House Publishers.

Johnson, S.L., and D.M. Pekelney. 1996. “Economic assessment of the Regional Clean
Air Incentives Market: A New Emissions Trading Program for Los Angeles.”
Land Economics 72: 277-97.

Jones, C. O. 1978. An Introduction to the Study of Public Policy, 2
nd
ed. Boston:
Duxbury Press.

Joskow, P.L., R. Schmalensee, and E. Bailey. 1998. "The Market for Sulfur Dioxide
Emissions." The American Economic Review 88: 669-85.

187
Kamieniecki, S., G. Gonzales, and R.O. Vos, (Eds.). 1997. Flashpoints in
Environmental Policymaking: Controversies in Achieving Sustainability. Albany:
SUNY Press.

Kamieniecki, S., D. Shafie, and J. Silvers. 1999. “Forming Partnerships in
Environmental Policy: The Business of Emissions Trading in Clean Air
Management.” American Behavioral Scientist 43: 107-123.

Kerr, R.L., S.J. Anderson, and J. Jaksch. 2000. “Crosscutting Analysis of Trading
Programs.” Learning from Innovations in Environmental Protection Research
Paper 6. Washington, D.C.: National Academy of Public Administration.

Kilpatrick, H.E. 2003. “Policy Oriented Social Science Research, Measuring the
Benefits of.” In Jack Rabin (Ed.), The Encyclopedia of Public Administration and
Public Policy, Volume II. New York: Marcel Dekker, pp. 931 – 938.

King, G., R.O. Keohane, and S. Verba. 1994. Designing Social Inquiry: Scientific
Inference in Qualitative Research. Princeton, NJ: Princeton University Press.

Kingdon, J.W. 1995. Agendas, Alternatives, and Public Policies, 2d ed. New York:
Harper Collins.

Kosobud, R.F. (Ed.). 2000. Emissions Trading: Environmental Policy’s New Approach.
New York: John Wiley & Sons.

Knaap, G.J. and T.J. Kim. 1998. “Environmental Program Evaluation: Framing the
Subject, Identifying Issues.” In G. J. Knaap and T. J. Kim Environmental
Program Evaluation: A Primer. Urbana, IL: University of Illinois Press, pp. 1-20.

Kraft, M. E. 1998. “Using Environmental Program Evaluation: Politics, Knowledge, and
Policy Change.” In G. J. Knaap and T. J. Kim (Eds.), Environmental Program
Evaluation: A Primer. Urbana, IL: University of Illinois Press. pp. 293-320.

Kraft, M.E. and N.J. Vig. 2000. “Environmental Policy from the 1970’s to 2000.” In
N.J. Vig and M.E. Kraft, Environmental Policy, 4
th
ed. Washington, DC: CQ
Press.

Kraul, C. 1997. “Companies Trade Smog Credits on Online Exchange.” Los Angeles
Times, April 30, p. D1.

Lasswell, H.D. 1971. A Preview of Policy Sciences. New York: American Elsevier.

188
Lester, J. P. and M.L. Goggin. 1998. “Back to the Future: The Rediscovery of
Implementation Studies.” Policy Currents 3 (No. 3, September): 1-9.

Lipsky, M. 1971. Street-Level Bureaucracy. New York: The Russel Sage Foundation.

Lynn, L.E., Jr. 1987. Managing Public Policy. Boston: Little, Brown.

McLean, B.J. 1996. “The Evolution of Marketable Permits: The US Experience with
Sulfer Dioxide Allowance Trading.” In D. Anderson and M. Grubb (Eds.),
Controlling Carbon and Sulfer Dioxide: Joint Implementation and Trading
Initiatives. London: The Royal Institute of International Affairs.

Mazmanian, D.A., and P.A. Sabatier. 1983. Implementation and Public Policy. New
York: Plenum Press.

Mazmanian, D.A., and Kraft (Eds.). 1999. Toward Sustainable Communities:
Transitions and Transformations in Environmental Policy. Cambridge: MIT
Press.

Meir, K.J. 1999. “Are We Sure Lasswell Did It This Way?” Policy Currents 9 (No. 1,
March): 5-8.

Moe, T.M. 1991. “Politics and the theory of organization. The Journal of Law,
Economics & Organization 7: 106-129.

OECD. 2002. Implementing Domestic Tradeable Permits: Recent Developments and
Future Challenges. Paris: OECD.

Ostrom, E. 1990. Governing the Commons. New York: Cambridge University Press

O’Toole, L.J., Jr. 2000. “Research on Policy Implementation: Assessment and
Prospects.” Journal of Public Administration Research and Theory 92: 1-22.

Pigou, A.C. 1920. The Economics of Welfare. London: Macmillan.

Polakovic, G. 2000, August 30. “DWP to Pay Pollution Penalty of $14 Million.” Los
Angeles Times, p. B1.

---------. 2001, April 17. “Innovative Smog Plan Makes Little Progress.” Los Angeles
Times, p. B1.

189
Pollesctsky, M. 1995. “Will a Market in Air Pollution Clean the Nation’s Dirtiest Air?
A Study of the South Coast Air Quality Management District’s Regional Clean
Air Incentives Market.” Ecology Law Quarterly 22: 359-411.

Portnoy, P. R. 1990. “Air Pollution Policy.” In P. R. Portnoy (Ed.), Public Policies for
Environmental Protection. Washington, DC: Resources for the Future, pp. 27-96.

Portnoy, P.R. and R.N. Stavins (Eds.). 2000. Public Policies for Environmental
Protection. Washington, DC: Resources for the Future.

Posner, R. 1992. Economic Analysis of Law. Boston: Little, Brown.

Ragin, C. C. 1987. The Comparative Method. Berkeley: University of California Press.

Rampedi, M.A. 2003. Implementing Adult Education Policy in the Limpopo Province of
South Africa: Ideal, Challenges, and Opportunities. Ph. D. diss, University of
Gronigen, Netherlands.

Reinhardt, F.L. 2000. Down to Earth: Applying Business Principles to Environmental
Management. Boston: Harvard Business School Press.

Ringquist, E.S. 1998. Efficiency Versus Equity in Environmental Protection: Trading
SO2 Emissions Under the 1990 Clean Air Act. Paper presented at the Annual
Meeting of the American Political Science Association, Boston, MA.

Rosenbaum, W.A. 1998. Environmental Politics and Policy, 4
th
ed. Washington, DC:
CQ Press.

Sabatier, P.A. 1986. “Top-down and bottom-up approaches to implementation research:
A critical analysis and a suggested synthesis.” Journal of Public Policy 6 (1): 21-
28.

Sabatier, P.A., and H.C. Jenkins-Smith. 1993. Policy Change and Learning: An
Advocacy Coalition Approach. Boulder, CO: Westview Press.

Sabatier, P.A., and D.A. Mazmanian. 1980. “The implementation of public policy: A
framework for analysis.” Policy Studies Journal 8: 538-60.

Sabatier, P.A., and D.A. Mazmanian. 1983. Can Regulation Work? The Implementation
of the 1972 California Coastal Initiative. New York: Plenum Press.

Shafie, D. 2002. Regulatory Agenda-Setting and Environmental Risk. Ph. D. diss,
University of Southern California.

190

Savas, E. S. 1987. Privatization: The Key to Better Government. Chatham, NJ;
Chatham House.

Scheberle, D. 1997. Federalism and Environmental Policy: Trust and the Politics of
Implementation. Washington, D.C.: Georgetown University Press.

Schneider, A.L. 1999. “Terminator! Who Me? Some Thoughts on the Study of Public
Policy Implementation.” Policy Currents 9: 1-5.

Schneider, A.L., and H. Ingram. 1997. Policy Design for Democracy. Lawrence:
University of Kansas Press.

Siebert, H.S. 1998. Economics of the Environment: Theory and Policy, 5
th
ed. New
York: Springer.

Smith, Z. 2000. The Environmental Policy Paradox, 3
rd
ed. Upper Saddle River, NJ:
Prentice Hall.

--------. 2004. The Environmental Policy Paradox, 4
th
ed. Upper Saddle River, NJ:
Prentice Hall.

South Coast Air Quality Management District (SCAQMD). 1991. Air Quality
Management Plan. Diamond Bar, CA: Author.

-----. 1994. RECLAIM: Program Summary. Diamond Bar, CA: Author.

-----. 1996. Annual RECLAIM Audit Report for the 1994 Compliance Year. Diamond
Bar, CA: Author.

-----. 1997. Annual RECLAIM Audit Report for the 1995 Compliance Year. Diamond
Bar, CA: Author.

-----. 1998. RECLAIM Program three-year Summary. Diamond Bar, CA: Author.

-----. 1999. Annual RECLAIM Audit Report for the 1997 Compliance Year. Diamond
Bar, CA: Author.

-----. 2000. Annual RECLAIM Audit Report for the 1998 Compliance Year. Diamond
Bar, CA: Author.

-----. 2001. Annual RECLAIM Audit Report for the 1999 Compliance Year. Diamond
Bar, CA: Author.

191

-----. 2002. Annual RECLAIM Audit Report for the 2000 Compliance Year. Diamond
Bar, CA: Author.

-----. 2003. Annual RECLAIM Audit Report for the 2001 Compliance Year. Diamond
Bar, CA: Author.

-----. 2004. Annual RECLAIM Audit Report for the 2002 Compliance Year. Diamond
Bar, CA: Author.

-----. 2005. Annual RECLAIM Audit Report for the 2003 Compliance Year. Diamond
Bar, CA: Author.

-----. 2006. Annual RECLAIM Audit Report for the 2004 Compliance Year. Diamond
Bar, CA: Author.

-----. 2007. Over a Dozen Years of RECLAIM Implementation: Key Lessons Learned
in California’s First Air Pollution Cap-and-Trade Program. Diamond Bar, CA:
Author.

Stephen S. 2002. “Ex-post evaluations of tradeable permits programmes.” In OECD,
Implementing Domestic Tradeable Permits: Recent Developments and Future
Challenges. Paris: OECD.

Stavins, R. N. 1998. "What Can We Learn from the Grand Policy Experiment? Lessons
from SO2 Allowance Trading." Journal of Economic Perspectives 12: 69-88.

-----. 2000a. "Market Based Environmental Policies." In P. R. Portnoy and R. N. Stavins
(Eds.), Public Policies for Environmental Protection. Washington DC: Resources
for the Future.

-------. 2000b. “Experience with Market-Based Environmental Policy Instruments.” RFF
Discussion Paper 00-009, Washington DC: Resources for the Future.

Thompson, D. B. 2000. “Political Obstacles to the Implementation of Emissions
Markets: Lessons from RECLAIM.” Natural Resources Journal 40: 645-694.

Tietenberg, T.H. 1985. Emissions Trading. Washington, DC: Resources for the Future.

------. 2000. Environmental and Natural Resource Economics, 5th ed. New York:
Addison-Wesley.

192
Vig, N.J., and M. Kraft. 2000. Environmental Policy, 4
th
ed. Washington, DC: CQ
Press.

Weimer, D.L., and A.R. Vining. 1989. Policy analysis: Concepts and practice, 2nd ed.
Englewood Cliffs, NJ: Prentice Hall.

Weiss, C. 1972. Evaluation Research: Methods for Assessing Program Effectiveness.
Englewood Cliffs, N.J.: Prentice Hall.

Wildavsky, A., and J.L. Pressman. 1973. Implementation. Berkeley: University of
California Press.

Wildavsky, A, J.L. Pressman, and G. Majone. 1983. Implementation, 2
nd
ed. Berkeley:
University of California Press.

Yin, R.K. 1994. Case Study Research, 2
nd
ed. Thousand Oaks, CA: Sage Publications.

Asset Metadata

Creator Wright, Christopher J. (author)

Core Title Assessing the implementation of the RECLAIM cap-and-trade market for pollution: measurement issues in counterfactuals, goal attainment, and command-and-control alternatives

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Political Science

Publication Date 12/01/2008

Defense Date 10/17/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Air pollution,cap-and-trade,command-and-control,environmental markets,environmental policy,Markets,OAI-PMH Harvest,policy evaluation,pollution markets,Public Administration

Language English

Advisor Barnes, John E. (committee chair), Crigler, Ann N. (committee member), Mazmanian, Daniel A. (committee member)

Creator Email wrightch@uamont.edu,wrightchristopherj@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1853

Unique identifier UC1228283

Identifier etd-Wright-2495 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-135695 (legacy record id),usctheses-m1853 (legacy record id)

Legacy Identifier etd-Wright-2495.pdf

Dmrecord 135695

Document Type Dissertation

Rights Wright, Christopher J.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract The creation of cap-and-trade (CAT) markets for pollution has long been suggested as an alternative to command-and-control (CAC) regulation. Economic theory suggests that CAT markets could achieve similar environmental outcomes to CAC in a more efficient manner. In addition, CAT holds the promise that they may deliver environmental improvements above and beyond CAC. CAT should also spur innovation in abatement technologies and processes. This dissertation uses the Regional Clean Air Incentives Market (RECLAIM) as a critical case study of the effectiveness of cap-and-trade programs. The main focus of this dissertation is in assessing the effects of the RECLAIM program on emissions levels. The main contribution of this dissertation is that it provides critical insight into measurement issues surrounding CAT markets. Most previous assessment efforts have used the level of tradable permits to assess policy effectiveness of emissions reductions. This dissertation argues that permit levels may not truly represent a CAC counterfactual. A better measure might be to compare emissions levels of regulated entities under CAT markets to those of CAC in a similarly situated environment. Using this latter measure this study finds that the RECLAIM program did not lower emissions levels below those found under CAC. Efficiency gains under RECLAIM may also have been overstated. Further, RECLAIM probably did not produce innovation. An implementation framework is adopted to explain why RECLAIM was never a pure example of a CAT market. As the policy has matured it has deviated farther and farther from the CAT ideal. Better measurement standards and more in depth assessments should produce better inferences on the extent to which CAT markets work under real world conditions. As CAT markets are becoming widely diffused, usable knowledge about their potential and limitations is critical.