RESEARCH

Industrial Ecology in Practice

The Evolution of Interdependence
at Kalundborg
John Ehrenfeld
Nicholas Gertler*
Technology Business and Environment Program
Massachusetts Institute of Technology
Cambridge, Massachusetts, USA

Keywords
eco-industrial park
green twinning
industrial ecosystems
industrial symbiosis
islands of sustainability
Kalundborg

I
Summary

The exchange of wastes, by-products, and energy among

closely situated firms is one of the distinctive features of
the applications of industrial ecological principles.This at=
ticle examines the industrial district at Kalundborg, Denmark, often labeled as an industrial ecosystem or
industrial symbiosis because ofthe many links among the
firms.The forces that led to its evolution and to the interdependencies are described and analyzed. Key has been a
sequence o f independent, economically driven actions.
Other potential forms of industrial linkages are critically
reviewed in the light of the Kalundborg experience.The
evolutionary pattern followed at Kalundborg may not be
easily transferable to greenfield developments.

Address correspondence to:

JohnEhrenfeld
Technology Business 61the Environment
Program
MassachusettsInstitute of Technology
CTPID, E40-421

Cambridge, MA 02139
USA
[email protected]

0Copyright 1997 by the Massachusetts

Institute of Technolcgy and Yale University
Volume 1, Number 1

*Nicholas Gertler is currently at Harvard Law School, Cambridge, Massachusetts, USA

journal of Industrial Ecology

67

RESEARCH

Introduction
Industrial ecology is a new concept emerging
in the evolution of environmental management
paradigms (Ehrenfeld 1995), and springs from interests in integrating notions of sustainability
into environmental and economic systems
(Allenby 1992; Jelinski et al. 1992; Allen and
Behmanish 1994; Ehrenfeld 1995). Environmental thinking has recently focused on a consciousness of the intimate and critical relationships
between human actions and the natural world,
and reflects limits in the current reliance on
command-and-control regulation in much of the
industrialized world. The critical problem is that,
for the most part, the economy operates as an
open system, drawing raw materials from the environment and returning vast amounts of unused
by-products in the form of pollution and waste.
The products that firms market are only a small
portion of what their processes turn out; a significant portion of their output eventually leaves
the economy as waste and returns to the environment in forms that may stress it unacceptably.
As long as attention is limited to products and
processes viewed in isolation, larger systemic
problems, such as the accumulation of persistent
toxic materials, will not be addressed.
Increased economic output will still cause increased environmental harm in such a frame of
analysis. Strong links between environment and
development emerged from the global consensus
following the 1992 Rio Earth Summit. For example, the recent report of the PresidentsCouncil for Sustainable Development (PCSD) in the
United States concludes, In the end, we found
agreement around the idea that to achieve our vision of sustainability some things must growjobs, productivity, wages, profits, capital and
savings, information, knowledge, educationand others-pollution, waste, poverty, energy
and material use per unit of output-must not
(PCSD 1996).
Accomplishing economic growth and environmental protection simultaneously requires
fundamentally new ways of examining and designing socioeconomic systems. One way to get
beyond the analytic limits of standard economic
theory is to draw on an ecological metaphor as a
means to better understand energy and material
68

Journal of Industrial Ecology

flows and as a guide to the design of industrial

structures and public policies (Daly 1991). Robert
Ayres has called systemwide material flows the
industrial metabolism of a n economy (Ayres
1989;Ayres and Simonis 1994). Emerging models
for operationalizing industrial ecology suggest
simple principles for design, for example, closing
material loops, avoidance of upsets to the metabolism of the natural system (toxics elimination and
pollution prevention), dematerialization, and
thermodynamically efficient energy utilization
(Ehrenfeld 1995; Lowe 1994; Tibbs 1992). These
principles have the potential of moving society
away from unsustainable development patterns by
greatly reducing the flows of energy and materials
into and out of an economy.
Moving from linear throughput to closed-loop
material and energy use are key themes in industrial ecology. Industrial activity based on such an
ecological conception can greatly reduce harmful
impacts associated with pollution and waste disposal, while easing the drain on finite strategic
resources. Familiar practices such as reuse,
remanufacture, and recycling represent a move in
this direction. Industrial symbiosis is closely related and involves the creation of linkages between firms to raise the efficiency, measured at
the scale of the system as a whole, of material and
energy flows through the entire cluster of processes. Some of the firms, viewed independently,
may appear to be inefficient compared to conventional measures of environmental performance. Yet environmental performance can be
superior in the overall group of firms because of
the linkages. The cascading use of energy and the
use of industrial by-products as feedstocks for processes other than the ones that created them is
fundamental to this approach. In such cases, byproducts can replace virgin materials as feedstocks. Energy cascading involves the use of the
residual energy in liquids or steam emanating
from one process to provide heating, cooling, or
pressure for another process. The evolution of a
set of interrelated symbiotic links among groups
of firms in an area gives rise to a complex that we
(and others) call an industrial ecosystem. This article examines an example of the development of
such an industrial eccxystem and the context in
which it emerged, identifying characteristics that
may be useful in policy design.

RESEARCH

Stable ecological systems are steady-state, entropy-minimizing, highly interdependent collections of producers and consumers (Prigogine
1955). These characteristics are different from
those of an economy modeled after standard
(economic) premises-quasi-equilibrium systems
of essentially independent entities. Observations
of materials flows and energy consumption in industrialized economies indicate highly dissipative
usage which translates into entropy-increasing
processes (Ayres 1994). Ayres notes that ecological systems are dissipative, also. At steady state,
however, they correspond thermodynamically to
conditions of minimum levels of entropy production (Prigogine 1955). Prigogine notes further
that such entropy-minimizing states in stable biological systems are accompanied by increases in
the interdependenceamong the entities. The notion of food webs, including detritivores (the
scavengers that consume the wastes of other species) as important members, highlights the idea
of closed or nearly closed material loops in such
stable systems.
The relationship between steady-state economies and thermodynamics was first elaborated
by Georgescu-Roegen (1971) and further used as
one of the key foundations in Herman Dalys
steady-state economic framework (Daly 1991).
Cloud (1977, 679) noted that, materials and
energy are the interdependent feedstocks of economic systems, and thermodynamics is their
moderator. Interestingly, he refers to the industrial ecosystem in the article, perhaps the
earliest use of this term. The importance of entropy is that it is a measure related to the practical availability of materials and energy in a
system (Daly 1991). Entropy, as a measure of disorder in a system, always increases as energy is
made available from its chemical potential in
fossil fuels and wastes are dissipated in the environment. Regaining the utility of the energy and
materials requires reversing the entropic flows
which can be done only at the expense of using
even more energy. Thus, economic arrangements that minimize the production of entropy
have some long-term advantages in the context
of sustainability, over and above arguments
based solely on efficiency.
All of this is prelude to introducing industrial
symbiosis, another ecological metaphor used to

describe a practical economic arrangement.

Symbiosis is a biological term referring to a
close sustained living together of two species or
kinds of organisms (Encyclopaedia Britannica
1992, 14: 1034). The term was used as early as
1873 by a German botanist, H. A. De Bary, to
describe the intimate, mutually beneficial coupling of fungi and algae in lichens. Symbiosis in
economic systems is manifest in the exchange of
materials and energy between individual firms
located in close proximity. It is a specific example of combining several generic industrial
ecological behavior patterns, for example, loop
closing and energy cascades (Ehrenfeld 1995).
Although not a necessary condition for all loopclosing exchanges, in particular those for commodity materials like scrap paper or steel,
proximity is a hallmark of industrial symbiosis.
This article argues that symbioses are distinct
and evolve through a different process from
these more common market activities.

The Kalundborg Industrial

Ecosystem
A highly evolved industrial ecosystem is located in the seaside industrial town of Kalundborg,
Denmark (Gertler and Ehrenfeld 1996). Eleven
physical linkages comprise much of the tangible
aspect of industrial symbiosis in Kalundborg (see
figure 1). The towns four main industriesAsnaes Power Station, a 1,500-megawatt coalfired power plant; a large oil refinery operated by
Statoil; Novo Nordisk, a maker of pharmaceuticals and enzymes; and Gyproc, a plasterboard
manufacturer-and several users within the municipality trade and make use of waste streams
and energy resources, and turn by-products into
raw materials. Firms outside the area also participate as recipients of by-product-to-raw-material
exchanges. The symbioses evolved gradually (see
table 1) and without a grand design over the past
25 years, as the firms sought to make economic
use of their by-products and to minimize the cost
of compliance with new, ever-stricter environmental regulations.
At the heart of this system of arrangements is
the Asnaes Power Station, the largest power
plant in Denmark. By exporting part of the formerly wasted energy, Asnaes has reduced the

Ehrenfeld and Gertfer, The Kalundborg Industrial Ecosystem

69

r
1

RESEARCH

Kemira
Acid Plant
(Jutland)

Sulfur

Waste

HeaA

Statoil Refinery
I

Waste

Greenhouses
(Discontinued)

District
Heating

Gyproc
Wallboard
Plant

Steam

Waste
Heat

r
Neighboring
Farms

Fish Farms

Novo Nordisk
Fertilizer

Figure I The industrial ecosystem at Kalundborg, Denmark

fraction of available energy directly discarded by

about 80%. Since 1981, the town of Kalundborg
has eliminated the use of 3,500 oil-fired residential furnaces by distributing heat from the power
plant through a network of underground pipes.
Homeowners pay for t h e piping, but receive
cheap, reliable heat in return. The power plant
also supplies heat to its own fish farm, exploiting
a n increase in growth rate from the warmer water. Sludge from the ponds is sold as fertilizer.
Asnaes also delivers process steam t o its
neighbors, N o v o Nordisk a n d Statoil. T h e
Statoil refinery receives 40% of its steam requirements while Novo Nordisk receives all of
its steam requirements from Asnaes. T h e decision to rely completely on Asnaes for steam was
made in 1982 when N o v o Nordisk was faced
with the need to upgrade and renovate its boilers. Buying steam from outside was seen as a
cheaper alternative. T h e two-mile-long steam

70

Journal of Industrial Ecology

pipeline built for the interchange paid for itself

in two years. In addition, thermal pollution of
t h e nearby fjord from the former Asnaes discharge has been reduced.
T h e power station also provides a gypsumcontaining feedstock to Gyproc, a neighboring
wallboard maker. In 1993, Asnaes completed the
installation of a $115 million sulfur dioxide
scrubber that produces calcium sulfate, or industrial gypsum, as a by-product a t a rate of 8085,000 tons per year. Conveniently, gypsum is the
primary ingredient of wallboard. Two-thirds of
Gyproc's gypsum needs are met by Asnaes's scrubber, while much of the rest comes from a scrubber
at a similar German power plant. Gyproc formerly obtained all its gypsum from Spanish openpit mines which still supply a small portion of its
needs. Fly ash and clinker, the remains of coalburning power generation, are sold by Asnaes for
road building and cement production.

RESEARCH

Table I Chronology of KalundboE development

~~

Year

Action

1959
1961

Asnae~power station commissioned

1972
1973
1976
1979
1981
1982
1987
1989
1990
1991
1992
1993

Statoil refinery commissioned; water

piped from Lake Tissa
Gyproc A / S built; gas piped from Statoil
Asnaes expands; draws water from
pipeline
Novo Nordisk begins shipping sludge to
farmers
Asnaes begins to sell fly ash to cement
producers
Asnaes produces heat for Kalundborg
Kommune
Asnaes delivers steam to Statoil and
Novo Nordisk
Statoil pipes cooling water to Asnaes
Novo Nordisk switches from Lake Tissa
to wells
Statoil sells molten sulfur to Kemira in
Jutland
Statoil sends treated waste water to
Asnaes for utility use
Statoil sends desulfirized waste gas to
Asnaes
Asnaes supplies gypsum to Gyproc

The Norwegian-owned Statoil refinery, producing a range of petroleum products from light
gas to heavy fuel oil, is located across the road
from Asnaes. Since 1972, Statoil has been piping the gas to Gyproc to fire wallboard drying
ovens, all but eliminating the common practice
of flaring waste gases. This fuel gas supplies all of
Gyprocs needs. For continuity, Gyproc installed
a butane backup system for those periods when
Statoil shuts down for maintenance. In 1990,
Statoil built a sour-gas desulfurization plant producing liquid sulfur that is promptly trucked
about 50 kilometers to Kemira where it is converted to sulfuric acid. With the sulfur removed,
Statoils gas is clean enough to be burned by
Asnaes as well.
Freshwater scarcity in Kalundborg has led to
water reuse schemes. Since 1987, Statoil has
piped 700,000 cubic meters per year of cooling
water to Asnaes, where it is purified and used as
boiler feed-water. Statoil has also made treated
waste water available to Asnaes, which uses
about 200,000 cubic meters a year for cleaning

purposes. Statoils investment in a biological

treatment facility produces an effluent sufficiently clean for Asnaess use. Symbiotic linkages
have reduced the water demand by around 25%.
A few miles from Asnaes and Statoil is Novo
Nordisk, a world leader in the production of insulin, enzymes, and penicillin. The plant employs about 1,000 people, roughly 10% of
Kalundborgs population. Novo Nordisk makes
its product mix by fermentation, based on agricultural crops that are converted to valuable
products by microorganisms. A nutrient-rich
sludge remains after the products are harvested.
Since 1976, Novo Nordisk has been distributing
the process sludge to about a thousand nearby
farms where it is spread on the land as fertilizer.
After heat treatment to kill remaining microorganisms, the sludge is distributed throughout the
countryside by a network of pipelines and tanker
trucks. Novo Nordisk produces 3,000 cubic
meters of sludge per day, but can only store three
days worth. The sludge is given away instead of
being sold, reflecting the firms concerns for disposal security. Three full-time employees coordinate its delivery. Distributing the sludge as
fertilizer was the least-cost way to comply with
regulations prohibiting Novo Nordisk from discharging the sludge directly into the sea.
Savings from more efficient utilization of resources and elimination of wastes are shown in
table 2.

Symbiosis in Other Settings

Full utilization of by-products is a concept
that has a long history in the petrochemical industry (Nemerow 1995). The Houston Ship
Channel is an oft-cited example of the application of industrial ecosystem concepts to such a
complex. The channel is lined with numerous
chemical, petrochemical, and energy-generating
facilities that have been using and exchanging
by-products for decades. In a typical exchange,
vent gas from an Amoco plant is rerouted to a
neighboring Chevron facility where it is compressed and separated into its hydrogen and propylene components. Prior to this arrangement,
the gas was burned in Amocos flare stack.
The availability of a by-product stream can
be a major factor in site selection in the petro-

Ehrenfefd and Gertler, The Kalundborg Industrial Ecosystem

71

RESEARCH

Table 2 Waste and resource savings at Kalundborg

Annual resource savings through interchanges

Water savings
Statoil-1.2 million cubic meters from Asnaes
(Novo Nordisk is now producing 900,000 cubic meters of treated water that is available to replace
freshsupplies, but is being discarded at present.)
Fuel savings
Asnaes-30,000 tons of coal (about 2% of throughput) by using Statoil fuel gas
About 19,000 tons of oil use by using fuel gas from Statoil in Novo Nordisks boilers and Gyproc
dryer fuel
Community heating via steam from Asnaes
Input chemicals
Fertilizer equivalent to Novo Nordisk sludge (about 800 tons nitrogen and 400 tons phosphorous)
2,800 tons sulfur
80,000 tons of gypsum
Wastes avoided through interchanges

200,000 tons fly ash and clinker from Asnaes (landfill)

80,000 tons scrubber sludge from Asnaes (landfill)
2,800 tons sulfur as hydrogen sulfide in flue gas from Statoil (air)
1 million cubic meters of water treatment sludge from Novo Nordisk (landfill or sea)
1,500-2,500tons of sulfur dioxide avoided by substituting coal and oil (air)
130,000 tons carbon dioxide avoided by substituting coal and oil (air)

chemical industry. The inherent flexibility of

petrochemical processes to accept a broad range
of feedstocks is an important factor leading to
widespread over-the-fence arrangements. Modern petrochemical plants can rearrange the molecules in a large variety of feed compositions to
obtain the desired product mix.
A Novo Nordisk plant in North Carolina
produces the same nutrient-rich sludge as its
Danish counterpart. As in Kalundborg, the
sludge is spread on farmland, providing fertilizer
for 5,000 acres. Gypsum from the scrubbers of
coal-fired power plants is used as raw material for
wallboard in a number of American states, including Texas and Florida. Wisconsin Public
Service, a public utility, is building a cogeneration power plant next to a paper mill in
Rhinelander, Wisconsin. The power plant will
burn wastes from Rhinelander Paper Company,
while providing all of the mills steam needs and
electricity for two local counties.
Industrial symbiosis is now receiving increasing attention worldwide. A major initiative has
arisen from the efforts of a Belgian economist
and entrepreneur, Gunter Pauli. His recognition
72

Journal of Industrial Ecology

of the limits of a single firm to reduce wastes to

zero, even with substantial efforts in pollution
prevention, gave rise to Paulis zero emissions
vision: multi-industry symbiotic clusters of factories. Pauli and others would extend the unplanned evolution process at Kalundborg to
create complexes of firms whose by-product
streams are linked. The Zero Emissions Research
Initiative, or ZERI, is an outgrowth of this
thinking. Based at United Nations University in
Tokyo, ZERI seeks to reduce a firms emissions
directly to the environment to zero while improving profitability. Based on the notion that
complementary industries can form zero-emissions clusters, one ZERI project aims at coupling
beer brewing with the raising of fish, mushrooms, and chickens. Through a carefully integrated coupling, ZERI predicts a sevenfold
increase in the total amount of nutrients available for human consumption and a fourfold increase in jobs, all on the basis of the standard
inputs to a beer brewery-cereals, yeast, and water (Pauli 1995).
One of the major policy thrusts of the report
of the U.S. Presidents Council for Sustainable

RESEARCH

Development is the creation of eco-industrial Each link in the system was negotiated, over a
parks where the firms would be closely linked period of some 25 years (see table I), as an indethrough waste and energy symbioses (PCSD pendent business deal and was established only
1996). As defined in the report, Eco-industrial if it was expected to be economically beneficial.
parks are an environmentally efficient version of Benefits are measured either as positive flows by
industrial parks. They follow a systems design in marketing a by-product (or obtaining feedstocks
which one facilitys wastes becomes another at prices below those for virgin materials) or as
facilitys feedstock, and they ensure that raw ma- savings relative to standard pollution control
terials are recycled or disposed of efficiently and approaches. This is the strength of the
safely. Specific projects are under way in Chat- Kalundborg approach: business leaders have
tanooga, Tennessee, Brownsville, Texas/ done the right thing for the environment in
Matamoros, Mexico; Baltimore, Maryland, and the pursuit of rational business interests. The
Cape Charles, Virginia. Each one has a different evolutionary nature can be interpreted as pointset of objectives and planning process (Gertler ing to a need to have both positive technical
1995). Chattanooga aims for the ZERl objective and economic factors appear simultaneously, a
of zero emissions. Baltimore is planned around condition that may be difficult or impossible to
four integrated strategies: loop closing and indus- realize in a forward-planning process.
trial ecology; network manufacturing; continu- Besides the basic chemical and other technical
ous improvement; and high-performance compatibility requirements of symbiotic partworkplaces and union-management cooperation ners, both need to recognize a net cost savings
(ETI 1995). Several planning guidelines for eco- relative to their options. The floor for economic
industrial parks have been recently published feasibility occurs when the difference in cost of
the by-product feed relative to virgin or other
(Lowe et al. 1995; Cote et al. 1994).
alternatives is less per unit throughput than the
cost of waste management to the producer. The
Patterns of Industrial
user can offer more than enough to the producer
Ecosystem Development
to offset the costs of waste treatment or disposal.
Although examples of closely linked or co- In practice the differential would also have to be
located symbioses as at Kalundborg are few, ma- large enough to account for transaction costs
terial exchange is a standard part of business and risks to both parties. Typical transaction
practice. The basic exchange of materials be- costs include regulatory, discovery, contracting,
tween manufacturers and their suppliers and cus- and monitoring costs. Discovery costs, the costs
tomers could be considered a kind of loose required to learn of the existence of an opportusymbiosis with benefits accruing to all parties. nity for exchange, can be high, and may be the
The recovery of scrap metals in many economies major impediment for material exchange of the
is another example of well-established industrial types discussed below. Brokerages and markets
symbiosis. Yet massive quantities of materials are serve to reduce these costs to the point that exroutinely discarded as wastes by industrial sys- change is economically rational. Exchange of
tems throughout the world. This section dis- material recovered from municipal waste streams
cusses a set of factors that both promote and (e.g., paper, metals, plastics) is now increasingly
inhibit the development of symbioses and indus- managed through commodity exchanges and
trial ecosystems.
electronic networks such as the Chicago Board
of Trade and the Global Recycling Network.
The buyer of by-products in a symbiosis takes
The Pattern of Economic Development
some risk by tying the firm to a single, outside
at Kalundbog
supplier and to the vagaries of supply continuity.
Many visitors come to Kalundborg looking The exchange of by-products and cascades of enfor the master plan. Despite its impressive re- ergy use, however, is not inherently different
sults, Kalundborg was not explicitly designed to from traditional supplier-customer relationships.
demonstrate the benefits of industrial symbiosis. Differential financial implications may be insig
Ehrenfeld and Gertler, The Kalundborg Industrial Ecosystem

73

RESEARCH

nificant. Provisions for standby supplies will add

cost. The seller also takes some risk given the possibility of upsets at the buyers facility that could
interrupt the outflow of the by-products. If this
were to happen, the by-products would instantly
become wastes to the seller and would need to be
disposed of according to the relevant regulatory
requirements.
Another economic factor is unrecognized environmental management savings to the firm
that could be realized by an exchange of materials. The emergence of full-cost accounting, a
form of activity-based accounting that more accurately identifies and allocates costs of complying with regulations and otherwise protecting
the environment, should enlarge the opportunities for symbioses and other forms of material exchange (Ditz, Ranganathan, and Banks 1995).
Future liability costs, for example, or the costs of
maintaining a legal staff for regulatory compli,
ance have been generally overlooked in evaluating innovative opportunities.
Organizational Arrangements
and Transaction Costs

In theory, symbioses can follow any of the

common types of industrial organization described, for example, by Williamson (1979) who
suggests that organizational arrangements between firms are shaped by efforts to minimize
transaction costs. Kalundborg is based on a complex of contracts and alliances that have arisen
with little or no institutional intervention. Unlike a spot market such as is typical in handling
metal scrap, this type of structure affords symbioses more certainty and continuity than exchanges in pure markets can offer. Vertical
integration, the common ownership of one or
more successive stages in the production process,
would go even further and might arise if continuity in the movement of by-products becomes a
critical factor. The cluster approach of ZERI appears to be heading in this direction.
Other forms of industrial organization more
common outside the United States have some
relevance to the emergence of symbiotic arrangements (Lenox 1995). The cross-ownership structure of the Japanese keiretsu is a highly elaborated
form of integration in which the transaction
74

journal of Industrial Ecology

costs and risks could be spread among potential

participants in an exchange of by-products. Another possibility is centralized ownership, as is
found in Germany where banks may own substantial equity in and participate actively in the
management of a number of firms. Other financial institutions could play a similar role. To reduce the coordinative problems with ecoparks
discussed above, some form of common ownership or institutional management power vested
in the developer of the park could improve the
context for the emergence of symbiotic patterns.
The eco-industrial park concept has many common characteristics with the more general notion of the manufacturing network form of
industrial development presented by Piore and
Sabel (1984) in their analysis of the success of
the artisan-based economy in the EmiliaRomagna region of Italy. Active trade associations; shared services, such as purchasing and
quality assurance; close family and community
ties are among the factors that contribute to the
success of such industrial districts.
Impediments other than strictly economic
ones exist as well, although Williamson might
argue all can be represented in terms of transactions costs. Symbiosis requires exchange of information about nearby industries and their
inputs and outputs that is often difficult or costly
to obtain. Kalundborgs small size of about
12,000 residents and its relative isolation have
made for a tight-knit community in which employees and managers interact socially with their
counterparts on a regular basis. This cultural feature leads to what a local leader calls a short
mental distance between firms (Christiansen
1994). Cultural pressures are also important. As
in many Scandinavian settings, there is a backdrop of environmental awareness.
In Kalundborg, n o deliberate institutional
mechanism was needed to promote conversations among the potential partners. Interfirm
trust is important in establishing alliances or
contracts among participants (Gulati 1995). A n
atmosphere of trust in Kalundborg existed even
in the absence of specific experience between
firms. In the United States, where there is a
strong tradition of company privacy, such natural communication is much more difficult to
find. The absence of a cultural context for ex-

RESEARCH

Early links at Kalundborg tended to involve

the sale of waste products without significant
pretreatment. This pattern includes the initial
sale of Statoils flue gas, Asnaess sale of fly ash,
clinker, waste heat and process steam, as well as
the use of cooling water to heat fish farm ponds.
These arrangements simply involved rerouting
of what was formerly waste, without any significant alteration. The more recent links, however,
have been created by and depend on the application of pollution control technologies. These
links, which comprise just over half of the interconnections, do not just move process by-products around. The processes and disposal
practices are controlled to make them more enTechnical Factors
vironmentally benign and, at the same time, to
In general, symbiotic industrial facilities render them more attractive as feedstocks. The
need to be in close proximity in order to avoid interposition of pollution control systems is imlarge transportation costs and energy degrada- portant in an industrial ecosystem as these techtion during transit. High-value by-products such nologies serve to concentrate dilute by-products
as pure sulfur from sour-gas treatment, such as into economically and technically attractive
that at Statoil, are exceptions. Contrary to the forms. The symbiotic relationships that comnotion of pollution prevention and zero waste at prise these links would not be attractive in the
a plant boundary, such as is, for example, the absence of such pollution control measures.
underlying policy goal of the U.S. Pollution Prevention Act of 1990, symbiosis may work best
Regulatory Context
when plants produce large quantities of waste.
This situation seems to be contrary to the noThe manager of the Asnaes plant believes
tion of eco-efficiency as applied to individual that existing economic incentives alone were
firms. It is not always best for either the bottom generally sufficient for much of the Kalundborg
line or the environment to reduce a single symbiosis (Christiansen 1994). Further symbiotic arrangements yielding environmental benplants waste to zero.
The cluster notion of ZERI works best, if efits are potentially available, but cost more
not requires, plants with large, continuous than conventional practices. Political impetus is
waste streams. Wastes that are largely organic necessary to go further, for example, requiring
in nature like the effluent from fermentation of emission reductions or adjusting prices to make
all sorts (e.g., pharmaceuticals or brewing), or symbiosis economically attractive. Yet such exraw agricultural or forestry wastes, are attrac- ternal signals alone are not sufficient, given that
tive as it is the organic carbon that is useful as innovative and pioneering cooperation is rea feedstock. Supply security is important to the quired among companies for symbiosis to occur.
user of the by-product stream exactly as would
The Danish regulatory framework has enbe the reliability of an otherwise virgin feed couraged the evolution of industrial symbiosis in
supply. Use of organic streams from fermenta- Kalundborg. Compared to the United States,
tion as feed or fertilizer requires little or no the Danish regulatory system is consultative,
matching of chemical characteristics other open, and flexible. Instead of being put on the
than assuring that toxic components or organ- defensive as is characteristic of a command-andisms are absent. Materials production, such as control framework, firms are required to be prothe manufacture of wallboard, is more techni- active by submitting plans to the overseeing
cally challenging and requires much closer county government detailing their efforts to
matching of compositions.
continually reduce their environmental impact.

change can be mitigated through institutional

mechanisms such as brokers or planning agencies. Such a brokers role is played by VP Resources of Clearlake, Texas, which defines its
role as more than a waste exchange (finding a
home for orphan chemicals) (Purcell 1995).
Where typical waste exchanges merely list available by-products, VP Resources performs every
function required to turn a by-product into a
feedstock, including finding appropriate uses,
dealing with regulatory agencies, brokering necessary agreements, and even transporting the
materials from the generator to the user.

Ehrenfeld and Gertler, The Kalundborg Industrial Ecosystem

75

RESEARCH

A dialogue then ensues in which the regulators

and the firm establish goals. A more flexible, cooperative relationship is fostered between government and the regulated industries, and as a
result, firms tend to focus their energies on finding creative ways to become more environmentally benign instead of fighting with regulators.
A key aspect of the flexibility is that regulatory
requirements are mainly in the form of performance standards stating the degree of the desired decrease, instead of technology standards
as is common in the United States. Technology
standards assure that uniformly effective pollution-control methods are adopted throughout a
given industry. They tend to hinder technological or infrastructural innovation, however
(Banks and Heaton 1995; Porter and van der
Linde 1995; Sparrow 1994). Many of the creative arrangements found in Kalundborg are
only possible where firms have flexibility in the
approaches employed to meet pollution-reduction targets. In the United States, little discretion is left to firms. There are disadvantages to
the Danish system, including potentially lower
levels of technical compliance and high transaction costs incurred in extensive consultations
around permitting. Although U.S. technology
standards are inflexible, they ensure a certain
minimum level of pollution control.
Regulatory requirements may preclude exchange or serve as very strong disincentives. In
the United States, for example, the Resource
Conservation and Recovery Act (RCRA) regulates the treatment and disposal of industrial
waste, but inadvertently impedes one of its objectives-conservation and recovery of resources. The statute is primarily concerned with
averting risks stemming from the improper management of hazardous waste. RCRA regulations
pursue this goal through a very extensive set of
specific, inflexible, and often confusing rules
governing the treatment, storage, and disposal of
industrial by-products (Hill 1991). RCRA regulations set forth specific detailed procedural and
technical requirements for the management of
an exhaustive list of particular types of waste
streams. With by-products being matched to a
particular, mandatory protocol, little room exists
for innovative schemes for their reuse as feedstocks elsewhere. This inflexibility is based in
76

journal of Industrial Ecology

larie-part on a deep-seated fear of sham recycling, which is an undertaking where the generator of a waste product makes a show of reusing
that by-product merely to escape treatment requirements (Comella 1993).Industrial symbiosis
must be distinguished from such efforts if it is to
develop within the current regulatory system.

Barriers and Limits to the

Development of Symbiotic
Cornmunities
The current vision of eco-industrial parks
closing material loops needs to be examined
carefully in light of the Kalundborg experience.
Opportunities for symbioses or clusters in the
sense of ZERI certainly exist, but not for more
than a small fraction of extant and planned industrial development zones in older economies
like the United States. The cluster approach appears more fitted to and is being promoted for
so-called greenfields or new initiatives in developing economies. Symbiosis, as in Kalundborg,
is not likely to become the main loop-closing
form for such complexes as it requires the presence of two or more firms that produce and consume a continuous stream containing useful
by-products. Such process industries are only a
small fraction of the typical manufacturingfirms
that comprise local industrial development areas. Many are small companies making parts or
assembling products, working metals and wood,
molding plastics, or cutting textiles. Finding
firms to colocate at the site of existing primary
sources of by-products, like breweries, would be
another form of this notion. Brownfields, as
abandoned or decaying industrial districts are
termed, are also candidates for the cluster
model. At the Second Annual World Congress
on Zero Emissions, Edgar Woolard, DuPonts recently retired CEO argued that such existing
sites should be developed before new greenfields
are taken out of their more natural uses (Krieger
1996). Developing economies with few
brownfields might disagree.
Companies that manufacture product components or assemble products produce two kinds
of wastes. One is essentially single material
wastes, such as unused or dirty solvents, paints,
or plastic or metal scrap. Such wastes can be

RESEARCH

used directly by other companies. According to

an informal survey done by MIT in 1994 (unpublished), waste exchanges acting as brokers
handle a very small fraction of such wastes. Unlike the continuous and relatively reliable production of process-industry by-products, these
manufacturing by-products have variable compositions and appear in small quantities. The
second waste type is mixed manufacturing
wastes including off-specification products.
Such wastes are similar to end-of-life durable
goods. To be useful, these materials must be either separated into their components or recycled
into an application that can use the bulk properties of the composite waste. Electronic scrap
might, for example, contain precious metals that
could be recovered and reused.
Exchanges of single material wastes lack the
technical and economic appeal of continuous
sources of feedstocks. Transaction costs for identifying and transferring such wastes are very high
and lot sizes are small. Loop closing by reusing
mixed wastes is unlikely to be viable in a localized area such as an ecopark. There is not
enough of any stream to provide for a freestanding, economically viable recovery activity. Very
high-value components such as precious metals
might lend themselves to closed loops, and there
might be opportunities for downcycling of
mixed plastics. Downcycling refers to material
recycling in products tolerating lower performance specifications than the original useplastic lumber, for example.
The critical missing factor is positive economical linkages as at Kalundborg. Recovery of
materials from mixed process waste is no different
from recovery from end-of-life products. Except
for precious metals, such recovered materials are
generally more expensive than virgin materials,
reversing the economic calculus. It would take
some form of public intervention, such as imposing large disposal costs or subsidies, to recovery
firms to create the favorable economics that led
to the evolution of Kalundborg. Taxes on pollution, waste disposal, and virgin materials (rather
than the current set of subsidies for many raw materials) would create incentives to keep more resources circulating in productive use, while
reducing the environmental impact of industrial
activity. But even with such policy intervention,

it may be difficult to produce conditions where

positive economics and environmental factors
arise simultaneously among a large number of
firms.
It was the arrival of such pairingdubbed by
some as green twinning4ver time that, arguably, produced the prerequisites for the development of the Kalundborg industrial ecosystem.
Together with the institutional context that
raised environmental values and enabled a high
information flow (and lower transaction costs),
the emergence of individual arrangements is the
expected outcome of a rational decision#making
process that should be reproduced in other similar contexts. Designing an industrial ecosystem
from the ground up is different and cannot follow the evolutionary path that contributed so
strongly to Kalundborgs positive development.
Simple probability suggests a very low chance of
finding pairs of coexisting positive environmental, technical, and economic factors among more
than one or two firms at any one time. If one assigns a binary value of zero (nonpositive linkage)
or one (positive linkage) to each of these three
factors with equal probability, basic combinatorial theory leads to an overall probability of only
1 in 64 that all three factors for both parties are
nonzero, such that the product is also nonzero,
that is, favorable. Adding more firms reduces the
joint probability even further. Interventionist
public or private initiatives can increase the
probability of positive values for each term, but
the overall probability may remain small.
The evolutionary path, on the other hand,
captures the opportunistic arrival of such joint
conditions, one case at a time. In this sense, the
evolution of a symbiotic industrial ecosystem is
little more than a special case of Williamsons
transaction cost model of the development of
patterns of industrial organization. Broad public
policy initiatives can nudge the process along by
creating the overall conditions that promote the
emergence of positive factors in individual cases
over time, but may falter where the promoters of
an eco-industrial park are attempting to create
all the linkages up front. An integrated and continuous planning process that explicitly seeks to
identify opportunities and barriers to symbiotic
development should increase the likelihood of
attracting firms willing to make the necessary ar-

Ehrenfeld and Gertler, The Kalundborg Industrial Ecosystem

77

RESEARCH

rangements. Such a middle-ground approach between pure laissez-faire and heavy-handed policy
intervention would seem to offer the best chance
of success to institutional developers.
A final barrier lies in t h e cognitive domain.
Wastes have such a long history of being ignored
that it is difficult for firms to integrate these outputs of their activities into their strategic processes. For firms t o t h i n k strategically and
systematically about their entire production
chain, they must change the mind-set that often
sees the world beyond the fence-line as only customers. This way of seeing the world is quickly
changing. New tools, such as life-cycle analysis
and new codes of practice like t h e Responsible
Care initiative of the chemical industry or t h e
IS0 14000 environmental management system
standards, are slowly introducing more systematic
and worldly models i n t o t h e underlying consciousness of firms and reducing the rigidity of traditional regulations (Freeman a n d Belcamino
1996). Together with new institutional approaches, these more deep-seated cultural changes
can provide a foundation from which symbioses
and other forms of material exchange begin to actually move economies toward sustainability.

References
Allen, D., and N. Behmanish. 1994. Wastes as raw
materials. In The greening of industrial ecosystems,
edited by B. R. Allenby and D. Richards. Washington, DC: National Academy Press.
Allenby, B. R. 1992. Achieving sustainable development through industrial ecology. International
Environmental Affairs 4( 1): 56-68.
Ayres, R. U. 1989. Industrial metabolism. In Technology and environment, edited by J. H. Ausubel and
H. E. Sladovich. Washington, DC: National
Academy Press.
Ayres, R. U. 1994. Industrial metabolism: theory and
policy. In Industrial metabolism: Restructuring for
sustainable development, edited by R. U. Ayres
and U. E. Simonis. Tokyo: United Nations University Press.
Ayres, R. U., and U. E. Simonis, eds. 1994. Industrial
metabolism: Restructuring for sustainable development. Tokyo: United Nations University Press.

Banks, R. D., and G. R. Heaton. 1995. An innovation-driven environmental policy. Issues in Science and Technology 12(1): 43-51.

78

journal of Industrial Ecology

Christiansen, V. 1994. Personal communication. May.

Cloud, P. 1977. Entropy, materials and posterity.
Geologische Rundschau 66(3): 678-696.
Comella, P. 1993. Understanding a sham: When is recycling, treatment? Environmental Affairs 20:
433-436.
C M , R., R. Ellison, J. Grant, J. Hall, P. Klynstra, M.
Martin, and P. Wade. 1994. Designing and operating industrial parks as ecosystems. School for
Resource and Environmental Studies, Dalhousie
University, Halifax, Nova Scotia.
Daly, H. E. 1991. Steady-state economics. 2d ed. Washington, DC: Island Press.
Ditz, D., J. Ranganathan, and R. D. Banks, eds. 1995.
Green ledgers: Case studies in corporate enuironmental accounting. Washington, DC: World Re-

sources Institute.
Ehrenfeld, John R. 1995. Industrial ecology: A strategic framework for product policy and other sustainable practices. In Green goods, edited by E.
Ryden and J. Strahl. Stockholm: Kretsloppdelegationen.
Encyclopaedia Britannica. 1992. The new Encyclopaedia
Britannica. Chicago: Encyclopaedia Britannica14:
1034.
ETI. 1995. Eco-industrial park progress report. Office of
Science and Technology Policy.
Freeman, D. J., and G. R. Belcamino. 1996. Brownfields redevelopment and I S 0 14000: A marriage
that makes sense. Corporate Environmental Strategy
3(2): 73-76.
Georgescu-Roegen, N. 1971. The entropy law and the
economic process. Cambridge: Harvard University Press.
Gertler, N. 1995. Industrial ecosystems: Developing
sustainable industrial structures. Masters thesis,
Massachusetts Institute of Technology.
Gertler, N., and J. R. Ehrenfeld. 1996. A down-toearth approach to clean production. Technology
Review 99(2): 48-54.
Gulati, R. 1995. Does familiarity breed trust?The implications of repeated ties for contractual choice
in alliances. Academy of Management Journal
38( 1): 85-1 12.
Hill, R. 1991. An overview of RCRA: The MindNumbing provisions of the most complicated
environmental statute. Environmental Law Reporter 21(May): 10255-10275.
Jelinski, L. W., T. E. Graedel, R. A. Laudise, D. W.
McCall, and C. K. Patel. 1992. Industrial ecology: Concepts and approaches. Proceedings of the
National Academy of Sciences 89 (February): 793797.
Krieger, J. H. 1996. Zero emissions gathers force as

RESEARCH

global environmental concept. Chemical 8Engt

neering News 74(28): 8-16.
Lenox, M. 1995. Barriers and opportunities for industrial symbiosis: A critical examination of altemative forms of industrial coordination. Unpublished
paper. Massachusetts lnstitute of Technology.
Lowe, E. 1994. Industrial ecology: Implications for
corporate strategy. Journal of Corporate Enuironmental Strategy 2( 1): 61-65.
Lowe, E. A., S. R. Moran, and D. B. Holmes. 1995.
Fieldbook for the development of eco-industrial
parks. Draft report. Oakland, CA: lndigo Development Company.
Nemerow, N. 1995. Zero pollution for industry. New
York: Wiley Interscience.
Pauli, Gunther. 1995. Personal communication. April.
PCSD (Presidents Council for Sustainable Development). 1996. Sustainabledeuehpment:A new consensus. Washington, DC.

Piore, M. J., and C. E Sabel. 1984. The second ituiustrial divide. New York: Basic Books.
Porter, M. E., and C. van der Linde. 1995. Green and
competitive: Ending the stalemate. Harvard
Business Review 73(5): 12C134.
Prigogine, I. 1955. Thermodynamics of irreversible processes. Springfield, IL: C. C. Thomas.
Purcell, V. 1995. Personal communication. February.
Sparrow, M. 1994. Imposing duties: Governments
changing approach to compliance. Westport, CT:
Praeger.
Tibbs, H. B. C. 1992. Industrial ecology-An agenda
for environmental management. Pollution Prevention Review 2(2): 167-180.
Williamson, 0. 1979. Transaction cost economics:
The governance of contractual relations. Journal
of Law and Economics 22: 233-262.

Ehrenfeld and Gertler, The Kalundborg Industrial Ecosystem

79

This article has been cited by:

1. Olli Salmi, Aino Toppinen. 2007. Embedding Science in Politics: Complex Utilization and
Industrial Ecology as Models of Natural Resource Use. Journal of Industrial Ecology 11:3, 93-111.
[Abstract] [PDF] [PDF Plus] [Supplementary material]
2. John R. Ehrenfeld. 2007. Would Industrial Ecology Exist without Sustainability in the Background?.
Journal of Industrial Ecology 11:1, 73-84. [Abstract] [PDF] [PDF Plus] [Supplementary material]
3. Marian R. Chertow. 2007. Uncovering Industrial Symbiosis. Journal of Industrial Ecology 11:1,
11-30. [Abstract] [PDF] [PDF Plus] [Supplementary material]
4. Noel Brings Jacobsen . 2006. Industrial Symbiosis in Kalundborg, Denmark: A Quantitative
Assessment of Economic and Environmental Aspects. Journal of Industrial Ecology 10:1-2, 239-255.
[Abstract] [PDF] [PDF Plus] [Supplementary material]
5. Tohru Morioka , Kiyotaka Tsunemi , Yugo Yamamoto , Helmut Yabar , Noboru Yoshida . 2005.
Eco-efficiency of Advanced Loop-closing Systems for Vehicles and Household Appliances in Hyogo
Eco-town. Journal of Industrial Ecology 9:4, 205-221. [Abstract] [PDF] [PDF Plus]
6. Stephen H. Levine . 2003. Comparing Products and Production in Ecological and Industrial Systems.
Journal of Industrial Ecology 7:2, 33-42. [Abstract] [PDF] [PDF Plus] [Supplementary material]
7. Catherine Hardy , Thomas E. Graedel . 2002. Industrial Ecosystems as Food Webs. Journal of
Industrial Ecology 6:1, 29-38. [Abstract] [PDF] [PDF Plus] [Supplementary material]
8. Marquita Hill , Thomas Saviello , Stephen Groves . 2002. The Greening of a Pulp and Paper Mill:
International Paper's Androscoggin Mill, Jay, Maine. Journal of Industrial Ecology 6:1, 107-120.
[Abstract] [PDF] [PDF Plus]
9. Pierre Desrochers . 2001. Cities and Industrial Symbiosis: Some Historical Perspectives and Policy
Implications. Journal of Industrial Ecology 5:4, 29-44. [Abstract] [PDF] [PDF Plus]
10. Derya B. zyurt , Matthew J. Realff . 2001. Combining a Geographical Information System and
Process Engineering to Design an Agricultural-Industrial Ecosystem. Journal of Industrial Ecology
5:3, 13-31. [Abstract] [PDF] [PDF Plus]
11. Michael T. Rock , David P. Angel , Tubagus Feridhanusetyawan . 1999. Industrial Ecology and Clean
Development in East Asia. Journal of Industrial Ecology 3:4, 29-42. [Abstract] [PDF] [PDF Plus]
12. Sara E. Keckler , David T. Allen . 1998. Material Reuse Modeling: A Case Study of Water Reuse,
in an Industrial Park. Journal of Industrial Ecology 2:4, 79-92. [Abstract] [PDF] [PDF Plus]
13. Daniel C. Esty, Michael E. Porter. 1998. Industrial Ecology and Competitiveness: Strategic
Implications for the Firm. Journal of Industrial Ecology 2:1, 35-43. [Abstract] [PDF] [PDF Plus]