- Email: [email protected]
On the joint determination of biological and economic systems
Ecological Economics 42 (2002) 301– 311 This article is also available online at: www.elsevier.com/locate/ecolecon
ANALYSIS
On the joint determination of biological and economic systems Chad Settle a,*, Thomas D. Crocker b, Jason F. Shogren b b
a Department of Economics, Uni6ersity of Tulsa, Tulsa, OK 74104 -3189, USA Department of Economics and Finance, Uni6ersity of Wyoming, Laramie, WY 82071 -3985, USA
Received 2 November 2001; received in revised form 12 April 2002; accepted 12 April 2002
Abstract Scarce natural resources and our choices to protect or develop them make ecological and economic systems jointly determined—human choices affect nature; nature affects human choices. This essay considers whether a dynamic model that integrates details of an economic system and an ecosystem with explicit feedback links between them yields significantly different results than does ignoring these links. We focus on the case of exotic invaders that put native species at risk in Yellowstone National Park. The results suggest that integration does matter— in each scenario, cutthroat trout populations differ in both magnitude and survival rates, depending on whether feedback is allowed between the two systems. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Economics; Ecology; Integration; Trout; Yellowstone; Exotic species
1. Introduction Despite the common rifts between the two disciplines, economics and the biological sciences are rife with similarities. Both are disciplines of limits — how to deal with scarcity. Whether it is a human reaction to a limited budget and unlimited wants or a fish’s response to limited food and unlimited appetite for reproduction, species must deal with their limits. The limiting factors in both disciplines drive their research efforts. Yet failure to account for joint influences upon these limits in * Corresponding author. Tel.: +1-918-631-3157; fax: + 1918-631-3546 E-mail address: [email protected] (C. Settle).
economic and biological systems can cause inaccurate perceptions of how each system works and provide misleading policy recommendations (Dasgupta et al., 2000). Joint determination creates a sequence of natural and human actions and reactions, and a feedback loop is born. The disturbances in one system set off repercussions in the other system, and these repercussions feedback into the system where the disturbances originated. The issue of risks to threatened and endangered species provides a vivid example. Conservation biologists often maintain that thresholds of species endangerments are functions of the present signs, trends and distributions of species’ populations and their likely interactions with habitats —strictly a bio-
0921-8009/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 0 0 9 ( 0 2 ) 0 0 1 0 5 - 2
302
C. Settle et al. / Ecological Economics 42 (2002) 301–311
logical question. We hold that this perspective is too narrow. Economic circumstances affect the quality of habitat. The circumstances that matter include the relative returns to human users from alternative sites, the relative returns from alternative uses on a particular site, and human welfare. Sites with low relative returns in their ‘highest and best’ use are more likely to be left undisturbed. Moreover, the rich can better afford to set aside quality habitat. Species survival is determined by economic as well as by biological parameters. Lacking unequivocal evidence to the contrary, the validity of separate treatments of the two parameter sets ought to be systematically demonstrated rather than routinely invoked. Models in which economic systems affect ecosystem outcomes are abundant in the economic literature. Fishery and forestry management models incorporate economic function into ecosystem representations (Clark, 1976). In the fishery context, these models often include predator –prey relations, selective harvesting or multiple species and may even introduce a spatial component which humans can influence. But these models are missing an important point of integration —feedbacks. Allowing for fishing pressure or harvest effort in a model accounts for how economic agents can alter the ecosystem. Having fishing pressure or harvesting effort as a constant, however, does not account for how humans adapt to a change in the fishery. With constant fishing effort, as fish populations fall due to an array of biological considerations, the harvest of fish also falls. Integrating economic systems and ecosystems via fishing effort captures this initial change. What it does not capture is how a change in one system can lead to a change in behavior in the other. When the fish species declines, will fishing effort actually be constant? When the fish population declines, an array of economic factors can cause humans to shift their efforts from one fish species to another or from fishing to other activities. This shift in behavior could lead to a different ecosystem steady state than if no account were taken of these feedbacks. This essay evaluates what taking account of the evolution of the details of jointly determined natural and human systems implies for a key, but
heretofore empirically untreated, question in the management of environmental and natural resources. Can a dynamic model that integrates an economic system and an ecosystem by formulating the details of feedback links between the two systems yield significantly different results than does the standard practice of giving short shrift to these links? We show how accounting for feedback between humans and nature affects the predicted ecological impacts— the population of a native prey species— caused by an exotic invader in Yellowstone Lake in Yellowstone National Park, Wyoming. The results show how acknowledgement of feedbacks might alter the core propositions, procedures, and public policy implications of both ecology and economics.
2. An application: exotic invaders in Yellowstone Lake Organisms that move beyond their traditional natural range can have undesirable ecological and economic consequences. Science has documented numerous examples of exotic plants and animals causing monetary and non-monetary damages (see Williamson, 1996). Consider a few classic examples. Field bindweed (Con6ol6ulus ar6ensis) is estimated to cause over $40 million in crop damages in Kansas every year (FICMNEW, 1998). The zebra mussel (Dreissena polymorpha) in the Great Lakes has significantly diminished phytoplankton biomass and harmed man-made structures (MacIsaac, 1996). The Nile perch (Lakes niloticus) has caused extinction of native fish and water quality problems in Lake Victoria. We now confront a similar problem in Yellowstone Lake, Wyoming, with the invasion of exotic lake trout (Sal6elinus namaycush). Yellowstone Lake is one of the last great inland fisheries in the United States for the native Yellowstone cutthroat trout (Oncorhynchus clarki bou6ieri ). Cutthroat trout are popular with fishermen and many predators such as ospreys (Pandion halieatus), white pelicans (Pelecanus erythrorhynchus), river otter (Lutra canadensis), and grizzly bears (Ursus arctos). In 1994, however, an angler caught a lake trout in Yellowstone Lake. Lake trout are an
C. Settle et al. / Ecological Economics 42 (2002) 301–311
exotic species to Yellowstone Lake— they prey upon but do not replace cutthroat trout in the food web. Fisheries biologists fear the lake trout population will expand and cause a severe decline in the cutthroat population (see Ruzycki and Beauchamp, 1997). If left unchecked, some biologists have predicted that this voracious exotic species could reduce the catchable-size cutthroat population from 2.5 million to 250,000– 500,000 within the near future (Kaeding et al., 1995). Furthermore, the ecosystem impact does not end with fewer cutthroat trout; grizzly bears, ospreys, eagles, river otters, and the other 40 species that rely on cutthroat as part of their food supply are put at risk (Schullery and Varley, 1995; Varley and Schullery, 1995, 1996). Most visitors to Yellowstone might not recognize the impact on cutthroats, but they are concerned with the loss of the larger, more charismatic species like grizzly bears that define Yellowstone for many people (Cherry and Shogren, 2001). This species interdependence could affect tourism in Yellowstone National Park. People come to Yellowstone to fish, bird watch, and they hope, see a grizzly bear. Some visitors might reduce the frequency of their visits, or they might enjoy the park less because of the changes within the ecosystem. An ecosystem externality may occur—the impact of lake trout on cutthroat trout shifts ecosystem equilibria, and the new equilibria affect the utility of park visitors through fewer cutthroats and the consequences for their predator species (see Crocker and Tschirhart, 1992). The species links within the ecosystem affect the utility of park visitors, whether they fish or not. Reducing risks to cutthroat trout, maintaining overall ecosystem integrity, and retaining visitor enjoyment provide ample justification for Yellowstone National Park managers to reduce pressure on the cutthroat trout population. By gillnetting lake trout, the park managers show, by maintaining a viable population of cutthroat trout or a large enough cutthroat trout population to sustain Yellowstone Lake as a cutthroat trout fishery, that they favor cutthroat trout over lake trout. What if accounting for the specifics of feedbacks between the two systems with and with-
303
out gillnetting leads to a difference in anticipated cutthroat trout populations? Then not addressing these feedbacks explicitly could lead Park managers to devote either too few or too many resources to the cutthroat trout problem. The question is whether accounting for the feedbacks between the Park’s ecological and economic systems provides additional useful data to managers who are considering incurring the costs of gillnetting the lake trout population in Yellowstone Lake.
3. Modeling biological –human interaction Traditionally, the specifics of threats to species and ecosystems have been estimated using one approach— the ‘damage function’ (DF) approach (see Freeman, 1993). The DF approach assumes the economic system and the ecosystem affect each other in a one-sided way. A change in the economic system is viewed as only changing the pressure on the ecosystem (Vitousek et al., 1997), or a change in the ecosystem is viewed as only changing the economic system (Daily, 1997). The DF approach, therefore, does not address the idea of co-evolution— the two-way interactions between human and natural systems (see Daly 1968; Norgaard, 1981).When an ecosystem changes, people change their behavior, which in turn reshapes the ecosystem, and so on. Ecosystem changes alter human productivity in the economic system. People recognize the change in their productivity when using the ecosystem and they adapt to this change, either by adapting the environment or by adapting to the environment. When people adapt, they alter the pressure they put on the ecosystem, leading to further changes in the ecosystem. The cycle continues. Can an explicit accounting of the specifics of these feedback links between the two systems yield different policy-relevant results than does assuming that no joint determination occurs? Though the broad outlines of ecological– economic system reciprocities have been acknowledged, the acknowledgements usually lump together key parameters of one or both systems. The lumping hinders predictions and evaluations
304
C. Settle et al. / Ecological Economics 42 (2002) 301–311
of which ecosystem services are to be usurped by humans, which waste flows are to be allowed to enter ecosystems, and which ecosystem biota and physical attributes are to be maintained. One can neither discriminate among the key parameters nor identify exactly how they enter the systems. Arguably, the contributions of natural environments to human well-being depend as much upon the complex details of ecological and economic states as upon the gross relationships between ecological and economic systems. Those analyses capable of admitting both the empirical detail and the jointness of ecological and economic systems are few (see Ayres and Kneese, 1969; Isard, 1972; Amir, 1979; Crocker and Tschirhart, 1992). These general equilibrium treatments work only with steady states. More recently a few authors (Swallow, 1996; Perrings, 1998; Sohngen and Mendelsohn, 1998; Brown and Layton, 2001) have set the structural details of jointness in a dynamic context. Here we describe a model and its results for two integrated systems: the economic system in Yellowstone National Park and the ecosystem in and around Yellowstone Lake. We have constructed a dynamic modeling framework incorporating many of the details of the flows inside each system and specifics of the reciprocal flows between the two systems. Incorporating the particulars of these links makes the model better reflect tradeoffs facing managers. We examine whether integrating economic systems and ecosystems matters, and if so, to what degree, by comparing the modeling results with and without accounting for the structure of feedbacks between the two systems.
4. Simulation model While exploiting a detailed analytical model of the Yellowstone Lake economic system and ecosystem, we use STELLA 2.0 software to simulate the importance of joint determination of the two species, lake and cutthroat trouts (Settle and Shogren (2002), provide the full mathematical specification of the model). Given the interactions and feedback loops between predator and prey
species and between species and humans, the full STELLA specification looks like a ball of yarn. We reduce the clutter for the reader by separating the full model into three main components: (1) lake trout; (2) cutthroat trout; and (3) human interactions. Components (1) and (2) capture the basic biological links between the species in and around Yellowstone Lake. Component (3) describes the human factors, including the interactions between humans and the ecosystem. We parameterized each component using population and consumption estimates for each species derived from the existing literature on the biology and socio-economics of Yellowstone Lake.1 These parameter estimates were used to build the interface between the economic system and the ecosystem. Now consider each component.
4.1. Model component 1: lake trout Diagram 1 shows the lake trout population and its direct biological links. The box labeled Lake Trout is a state variable indicating the population of lake trout in Yellowstone Lake at any given time. Deaths (LT natural deaths) are a flow out of the state variable, a decrease in the lake trout population. Births (LT births) are a flow into the state variable, an increase in population. The births are capped by a density dependence factor (DDFlt), the growth of lake trout is shown with a compensation model— the growth rate of lake trout declines with increases in the population of lake trout. The density dependence factor relates births of lake trout to its primary food source— cutthroat trout. As the cutthroat trout population declines, fewer lake trout can be supported in the lake. Now lake trout reallocate more time/energy to find food and less time/energy to spawn, which reduces spawning success. The reduction in spawning success reduces the number of births of lake trout. Note that the population of cutthroat trout also plays a key role in the density depen1 This research includes Blanchard (1980), Davenport (1974), Diem and Pugesek (1994), Kaeding (1995), Kaeding et al. (1995), Ruzycki and Beauchamp (1997), Swenson (1979), Stapp and Hayward (1998, 1999), Varley (1983), Varley and Schullery (1995, 1996), Wuerthner (1995).
C. Settle et al. / Ecological Economics 42 (2002) 301–311
dence factor, which we now consider in more detail.
305
the density-dependence factor limits the births of the species and places a carrying capacity on the cutthroat trout population. (2)
(1)
4.2. Model component 2: cutthroat trout
4.3. Model component 3: human impact
Diagram 2 shows the cutthroat trout population and its biology. The box labeled Cutthroat Trout is a state variable representing its population in the lake. Again, the two flows affecting population are births and natural deaths of cutthroat trout. A key difference between the cutthroat and lake trout components is that we have three species feeding upon cutthroat trout — lake trout, birds of prey, and grizzly bears. Therefore, the causes of cutthroat trout deaths include natural causes (CT natural deaths), predation by lake trout (CT killed LT), predation by birds of prey (CT killed B), and predation by grizzly bears (CT killed G). Each relationship is a function of the cutthroat trout population and the predatory species populations (state variables for these predatory species included in diagram 2). Lake trout are not likely to replace cutthroat trout in the diet of these predators since lake trout primarily stay in deep water and spawn in the lake instead of in the streams, where predators such as grizzly bears catch cutthroat trout (Stapp and Hayward, 1999). In addition, a density-dependence factor for cutthroat trout enters the model; once again the population of cutthroat trout is shown with a compensation model. Similar to the lake trout,
Now consider the human interaction with the ecosystem. Diagram 3 shows this interaction as captured by a representative visitor to Yellowstone National Park and its park manager. First, consider the visitor. Our visitor gains both direct and indirect benefits from species when visiting Yellowstone. The direct effects include: (1) fishing for lake trout (Human LT catch) and cutthroat trout (CT catch), which depends on the population of the fish and the time spent fishing (T F); 2 and (2) visiting the core attractions like Old Faithful and other geothermal activity, which we define as a park public good (X).The indirect effects include seeing birds of prey (Birds 6iewed) and grizzly bears (G 6iewed) while driving or fishing. The number of bird and grizzly bear sightings depends on species populations and time spent fishing (T F) and driving (T X).
2 Visitors are not divided into those who fish for lake trout and those who fish for cutthroat trout since Yellowstone Lake is a cutthroat trout fishery. People who intend to catch lake trout during a visit are able to go to Jackson Lake or other lake trout fisheries in northwestern Wyoming.
306
C. Settle et al. / Ecological Economics 42 (2002) 301–311
(3)
Our second human interaction is through the Park managers. These managers decide how to allocate a fixed budget for the park. In our model, they spend this fixed budget3 ($T) on two activities —improving the park public good4 ($X), and gillnetting the exotic lake trout to reduce pressure on the native cutthroat trout ($LT). The success of the gillnetting program, the number of lake trout killed by gillnetting (LT killed NPS), depends on both the budget spent gillnetting and on the lake trout population. This human interaction with the ecosystem allows us to capture the feedback links between the economic system and the ecosystem for Yellowstone Lake. These feedback links are captured by how the introduction of lake trout into Yellowstone Lake leads to changes in the ecosystem in and around Yellowstone Lake. The representative visitor responds to these changes by altering his behavior. When lake trout reduce the population of cutthroat trout, he finds it more difficult to 3 This fixed budget is always spent. The budgetary difference under the scenarios is the relative amount of the budget spent gillnetting lake trout versus improving the park public good. This fixed budget is also assumed constant. The policy effect on visitation is outside the scope of this model. 4 We use the condition of roads in the park as a proxy for the park public good.
catch cutthroat trout. Since he has less fishing success—the shadow price of fishing increases relative to the prices of alternative activities, he reallocates his time away from fishing at the Lake and toward other activities inside or away from the park where productivities have not declined, e.g. visiting Old Faithful. Therefore, the total number of fishing hours at Yellowstone Lake falls as the cutthroat population declines. This dynamic feedback loop does not jump from one steady state to another. The visitor reduces his time fishing in reaction to the changing ecosystem. Since the ecosystem is changing for several years, humans are altering their behavior for several years as well. Also, when humans alter their behavior, they change the pressure put on the fish species, which changes the path toward a new steady state for those species. This dynamic process continues until a steady state has been reached for both the ecosystem and the economic system— in which all species, including humans, are no longer reacting to a change in the ecosystem. As a comparative baseline, we presume that without such dynamic feedback loops the visitor does not respond to changes in the ecosystem. When lake trout reduce the population of cutthroat trout, assume he acts in the same manner as he did before lake trout were present. By measuring whether including feedbacks in our integrated model makes a difference in predicted fish populations, we can address the central question in this paper— how far off might the predictions be if we do not address the feedback loops between the two systems?
5. Results We consider three cases, each with and without feedbacks. First is the best-case scenario in which lake trout are immediately eliminated from Yellowstone Lake without cost. Cutthroat trout return to the world they had before lake trout were introduced. While infeasible in reality, the bestcase scenario defines the upper baseline on our indicators of well-being. Second, the worst-case scenario occurs when lake trout are left alone
C. Settle et al. / Ecological Economics 42 (2002) 301–311
307
Fig. 1.
without any interference from the National Park Service. Park managers do not attempt to help cutthroat trout survive. Instead, lake trout and cutthroat trout are left to reach their own steady state equilibrium. This worst-case produces the lower benchmark since it produces the smallest cutthroat trout population, and the lowest odds of cutthroat trout survival. Third, our policy scenario has the National Park Service reducing the risk to cutthroat trout by gillnetting lake trout.5 Assume the Service’s current level of expenditures on gillnetting lake trout is continuous and perpetual. We use the population of cutthroat trout as a yardstick.6 If the population of cutthroat trout is similar with and without feedbacks, an accounting of any feedbacks between the Yellowstone Lake ecosystem and economic system might not be worth the effort. If the resulting populations differ significantly within all cases, the accounting is more likely to provide the foundation for an improved understanding of the behavior of both systems. Table 1 and Fig. 1 summarize the results for each scenario. Under the best-case scenario without feedbacks, the steady state population of cutthroat trout is about 2.7 million. With feedbacks, 5 Gillnetting is used as the policy alternative since it is the method used by the National Park Service, and research determining the costs of reducing the lake trout population under various alternatives found gillnetting to be the lowest budgetary cost method (Varley and Schullery, 1995). 6 Listed populations are the steady state populations. The time to approach the steady state populations varies by scenario, with most approaching the steady state within 100 –200 years.
Table 1 Resulting cutthroat trout populations with and without feedbacks
Best-case scenario Policy scenario Worst-case scenario
No feedbacks
Feedbacks
2,700,000 1,900,000 900,000
3,400,000 2,300,000 0
the steady state population is about 3.4 million. The difference arises from fishermen’s behaviors. Without feedback, fishermen continue to fish as before, putting constant pressure on the cutthroat. With feedback, fishermen exploiting declining cutthroat populations adapt by fishing less and visiting other attractions more. Reduced human pressure on the cutthroat allows its population to increase by an amount greater than with constant fishing pressure. The resulting population of cutthroat trout is therefore greater with feedback from human adaptation. The results are similar for the policy scenario. Without feedback, visitors spend a constant amount of time fishing and put a constant pressure on cutthroat trout. The steady state population of cutthroat trout is 1.9 million. With feedback, however, the cutthroat population increases to 2.3 million. Again accounting for feedback matters—predicted populations are 20% greater. When the cutthroat trout population declines, fishermen are less successful. They respond by fishing less for cutthroat trout, which in turn allows the cutthroat trout population to increase over what it would be under constant fishing pressure.
C. Settle et al. / Ecological Economics 42 (2002) 301–311
308
Table 2 How system shocks affect the steady state populations of cutthroat trout Visitation at 5 million
Best-case scenario Policy scenario Worst-case scenario
No feedbacks Feedbacks
No fishing
2,300,000
3,100,000
4,700,000
1,700,000
2,000,000
2,700,000
1,100,000
0
0
The worst-case scenario tells a similar story but with a policy conclusion opposite that for the best-case and the policy scenarios. Assuming the lake trout population increases unabated, the cutthroat trout population falls dramatically. Without feedback, visitors put constant pressure on the cutthroat trout, and its steady state population falls to 0.9 million. With feedback, visitors again reduce the pressure on cutthroat trout, which should increase its steady state population. But a powerful countervailing effect exists—when people fish less, the incidental catch of lake trout also declines. As in the best-case and the policy scenarios, when visitors fish less, the incidental catch of lake trout drops, which in turn allows the larger population of lake trout to kill more cutthroat trout. But in this worst-case, the effects of less fishing pressure on the lake trout population dominates its cutthroat effect, supporting Volterra’s principle (Braun, 1975). Now the steady state population of cutthroat trout drops to zero. What matters is that when people shift their time away from fishing as the cutthroat trout population declines and the lake trout population increases, the incidental catch for fishermen of lake trout becomes an important control of the lake trout population. People shift away from fishing and that important control of lake trout declines. While one usually thinks of fishing pressure to be a pressure on cutthroat trout, in this case the more critical pressure is on lake trout, which were a higher proportion of fishermen’s catch. A no-feedback model suggests a healthier outcome than might actually exist—nearly 1 mil-
lion cutthroat trout versus none is a significant difference however one measures it. Feedbacks yield both different magnitudes and survival rates, suggesting that one discounts the importance of feedback loops at one’s own risk. We see that the system yields different outcomes under the scenarios with and without feedbacks—now consider the robustness of this result to two external shocks to the environment: (1) more fishing pressure due to increased visitation to the park; and (2) no fishing pressure due to a complete ban on fishing. First, we implement the visitation shock by increasing visitation to the park from 3 to 5 million visitors, the estimated visitors in 50 years if current trends continue (2050).7 Table 2 shows that the same patterns emerge with the population shock. More visitation increases fishing pressure, which lowers steady state populations for cutthroat trout, but we still have a positive population of cutthroat trout in every scenario except the worst-case scenario with feedbacks. Under the best-case scenario, the steady state population of cutthroat trout is 2.3 and 3.1 millions without and with feedbacks. With the policy scenario, the steady state population of cutthroat trout is 1.7 and 2 millions without and with feedbacks. Finally, under the worst-case scenario, the steady state population of cutthroat trout is 1.1 million and 0 without and with feedbacks. Shocking the system by adding a lot more visitors does not alter the direction of the results, only the size of the steady state populations. Feedbacks still matter. The second shock to the system is to ban fishing at the Lake— visitors are prohibited from catching either lake trout or cutthroat trout. Without fishing pressure on the ecosystem, the 7 We estimated the anticipated level of visitation in 2050 by running a univariate time series model on National Park Service visitation data between 1970 and 2000, and then extrapolated the results to 2050. The data indicated the expected level of visitation at 2050 was between 4 and 6 million, depending on which date was used as the starting value. Estimated visitation based on data from 1970 to 2000 was about 5 million, from 1980 to 2000, visitation was about 6 million; and from 1990 to 2000, anticipated visitation was about 4 million. We use the midpoint of these three estimates, i.e. 5 million. Data was collected from Yellowstone National Park’s webpage (www.nps.gov/yell).
C. Settle et al. / Ecological Economics 42 (2002) 301–311
steady state population of cutthroat trout increases under both the best-case scenario and the policy scenario to 4.7 and 2.7 millions.8 Under the worst-case scenario, the steady state population of cutthroat trout falls to 0, once again highlighting the importance of the inadvertent catch by visitors of the predatory lake trout. Neither visitation nor fishing ban shocks reduce the importance of feedbacks in modeling economic system–ecosystem interactions.
6. Concluding remarks Mark Twain said ‘‘Pain there was none, nor infirmity, nor any physical signs to mark the flight of time … in Eden’’. But when ‘‘Eve reached for an apple!— oh farewell Eden, and your sinless joys, come poverty and pain, hunger and cold heartbreak, … strife, …, age, weariness, remorse …’’ (Twain, 1991). Our aspirations and our temptations have cursed us with scarcity. We yield to human arrogance and ignorance— reaching for and biting the apple, and in so doing, we affect nature, ‘oh farewell Eden’, and thus cause nature to affect us—‘come poverty and pain’. This curse of Adam and Eve says that scarcity and the choices it necessitates make ecological and economic systems jointly determined: human choices affect nature; nature affects human choices. The curse of scarcity matters in the management of exotic invaders in Yellowstone Lake. Our results suggest that a dynamic model that integrates biological and economic systems with explicit treatment of some of the complexities of feedback links between the two systems yields significantly different results than when one ignores these links. In each scenario, cutthroat trout populations differ in both magnitude and survival rates once feedback is allowed between the two systems. For the best-case and the policy scenarios, no feedback leads to a lower estimate of the steady state population of cutthroat trout than 8 Eliminating fishing as an option yields no differences between feedbacks and no feedbacks since the option to fish is gone— all visitors spend no time fishing once it has been prohibited.
309
does feedback. But with the worst-case scenario, ignoring feedback actually leads the model to a higher estimate of the cutthroat population than does the model with feedback— a steady state population of 1 million versus none. This suggests that basing policy recommendations in Yellowstone Lake on data from models without feedback puts cutthroats at greater risk than would be true if feedback was explicitly considered. The potential policy relevance of the joint determination of biological systems and economic systems and the feedback it implies extends well beyond the Yellowstone Lake case. Policy makers should take the curse of Adam and Eve seriously. Consider AIDS, for example. Traditional epidemiological practice treats a person’s decisions concerning frequency of sexual contact and number of partners to be independent of the prevalence of the disease. But concede that people make choices about contacts and partners based on the odds of infection and the benefits and costs of what they can do to influence these odds. The predicted incidence of AIDS will be misestimated if the determinants of people’s choices are not taken into account. The light cast by making joint determination visible raises features of the AIDS system which would otherwise go unnoticed. Routine dismissal of the joint determination could compel similar errors in valuing alternative environmental states. For example, the gains from reducing the risk of lead poisoning in U.S. children have been estimated to at least double when one accounts for parents’ decisions to reduce children’s exposures to and body burdens of lead (Agee and Crocker, 1996). The inclusion of the behavioral interactions of economics and ecosystems for protection of North Carolina brackish wetland acreages leads to an average ninefold increase in calculated benefits (Swallow, 1996). In each case, the added benefits result from combinations of protection expenditures and natural asset price increases avoided, and the effects upon natural asset productivity of human adaptations to these avoided expenditures and price increases. The routine acceptance of the separability of natural events and human adaptations is not just one move in the long line of trade-offs between tractability and completeness which effective sci-
310
C. Settle et al. / Ecological Economics 42 (2002) 301–311
ence in support of policy requires. The separability question runs much deeper. The habitual invocation of separability obstructs economists’ and other social scientists’ opportunity — and obligation — to help define the thresholds of human health and ecosystem resilience, which underpin environmental policy. Like the snake in the apple tree of Eden, the invocation does a serious disservice to the common weal. One has to go beyond recognizing that the snake exists—actively seeking out more ways to map the details of scarcity and the jointness of natural and human systems is one path. In their extant studies of environmental –economy interactions, the two disciplines are biased in opposite directions. Ecologists are prone to regard anthropogenic environmental impacts as exogenous parameter shocks; they disregard the complexity of economic systems. Economists tend to view the environment as an aggregate or macro form in which interactions with the form can be dismissed; they therefore disregard the complexity of the ecological system. Explicit recognition of joint determination has an additional advantage: both ecology and economics can be better motivated to account for the complexities of both ecosystems and economic systems.
Acknowledgements We thank Greg Hayward, Paul Stapp, Wayne Hubert, the Yellowstone National Park for cooperation, and the USEPA EPSCOR Grant c R826281-09-0 and Stroock Professorship for financial support.
References Agee, M., Crocker, T., 1996. Parental altruism and child lead exposure. Journal of Human Resources 31, 677 –691. Amir, S., 1979. Economic interpretation of equilibrium concepts in ecological systems. Journal of Social and Biological Structures 2, 293 –314. Ayres, R., Kneese, A., 1969. Production, consumption, and externalities. American Economic Review 59, 282 –297. Blanchard, B.M., 1980. Grizzly bear – habitat relationships in the Yellowstone area. International Conference of Bear Reservation and Management 5, 118 –123.
Braun, M., 1975. Differential Equations and Their Applications: An Introduction to Applied Mathematics. SpringerVerlag, New York, 718 p. Brown, G., Layton, D.F., 2001. A market solution for preserving biodiversity: the Black Rhino. In: Shogren, J.F., Tschirhart, J. (Eds.), Protecting Endangered Species in the United States: Biological Needs, Political Realities, Economic Choices. Cambridge University Press, New York, NY, pp. 32– 50. Cherry, T.L., Shogren, J.F., 2001. Invasive species management for the Yellowstone Lake ecosystem: what do visitors think? Yellowstone Science 9, 11 – 15. Clark, C., 1976. Mathematical Bioeconomics: the Optimal Management of Renewable Resources. Wiley, New York, 352 p. Crocker, T.D., Tschirhart, J., 1992. Ecosystems, externalities, and economies. Environmental and Resource Economics 2, 551 – 567. Daily, G. (Ed.), 1997. Nature’s Services: Societal Dependence on Natural Ecosystems. Island Press, Washington DC, 392 p. Daly, H., 1968. On economics as a life science. Journal of Political Economy 76, 392 – 406. Dasgupta, P., Levin, S., Subchenco, J., 2002. Economic pathways to sustainability. Bioscience 50, 339 – 345. Davenport, M.B., 1974. Piscivorous Avifauna on Yellowstone Lake, Yellowstone National Park. Completion report to Yellowstone National Park, Yellowstone Park, Wyoming, USA. Diem, K.L., Pugesek, B.H., 1994. American white pelicans at the Molly Islands, in Yellowstone National Park: twentytwo years of boom-and-bust breeding, 1966 – 1987. Colonial Waterbirds 17 (2), 130 – 145. Federal Interagency Committee for the Management of Noxious and Exotic Weeds (FICMNEW), 1998. Invasive plants: changing the landscape of America, Fact Book. FICMNEW, Washington DC. Freeman III, A.M., 1993. The Measurement of Environmental and Resource Values: Theory and Methods, Resources for the Future, Washington DC. Isard, W., 1972. Ecologic – Economic Analysis for Regional Development. The Free Press, New York, NY, 270 p. Kaeding, L.R., 1995. Annual Project Report for 1994. Fishery and Aquatic Management Program, Yellowstone National Park. Kaeding, L., Boltz, G., Carty, D., 1995. Lake trout discovered in Yellowstone Lake, The Yellowstone Lake crisis: confronting a lake trout invasion. In: Varley, J., Schullery, P. (Eds.), A Report to the Director of the National Park Service, Yellowstone Center for Resources National Park Service, Yellowstone National Park, Wyoming, pp. 4 – 11. MacIsaac, H., 1996. Potential abiotic and biotic impacts of zebra mussels on the Inland Waters of North America. Amercian Zoology 36, 287 – 299. Norgaard, R.B., 1981. Sociosystem and ecosystem coevolution in the Amazon. Journal of Environmental Economics and Management 8, 238 – 254.
C. Settle et al. / Ecological Economics 42 (2002) 301–311 Perrings, C., 1998. Resilience in the dynamics of economy – environment systems. Environmental and Resource Economics 11, 503 – 520. Ruzycki, J., Beauchamp, D., 1997. A bioenergetics modeling assessment of the lake trout impact in Yellowstone Lake. In: Hamre, R. (Ed.), Wild Trout IV, Trout Unlimited and Federation of Fly Fishers. Fort Collins, CO, pp. 127 – 133. Schullery, P., Varley, J., 1995. Cutthroat trout and the Yellowstone Lake ecosystem. In: Varley, J., Schullery, P. (Eds.), The Yellowstone Lake Crisis: Confronting a Lake Trout Invasion, A Report to the Director of the National Park Service, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, Wyoming, pp. 12 –21. Settle, C., Shogren, J., 2002. Accounting for Feedback between Biological and Economic Systems: The Case of Exotic-Native Conflicts in Yellowstone Lake. Working paper, University of Tulsa. Sohngen, B., Mendelsohn, R., 1998. Valuing the impact of large-scale change in a market: the effect of climate change on U.S. timber. American Economic Review 88, 686 –710. Stapp, P., Hayward, G.D., 1998. Estimates of Predator Consumption of Yellowstone Cutthroat Trout. Working paper, Department of Zoology, University of Wyoming. Stapp, P., Hayward, G.D., 1999. Effects of Introduced Lake Trout in Yellowstone Lake: Insights from Demographic Analysis of Yellowstone Cutthroat Trout. Working paper, Department of Zoology, University of Wyoming. Swallow, S., 1996. Resource capital theory and ecosystem economics: developing nonrenewable habitats with hetero-
311
geneous quality. Southern Economic Journal 63, 106 – 123. Swenson, J.E., 1979. Factors affecting status and reproduction of ospreys in Yellowstone National Park. Journal of Wildlife Management 43 (3), 595 – 601. Twain, M., 1991. The Bible according to Mark Twain. In: Baetzhold, H.Y., McCollough, J.B. (Eds.), Writings on Heaven, Eden, and the Flood, The University of Georgia Press, Athens, GA, pp. 53, 66. Varley, J.D., 1983. The use of restriction regulations in managing wild salmonoids in Yellowstone National Park, with particular reference to cutthroat trout. In: Proceedings of the Olympic Wild Fish Conference, Penadweuh College, Port Angeles, Washington, pp. 145 – 156. Varley, J., Schullery, P., 1995. Socioeconomic values associated with the Yellowstone Lake cutthroat troat. In: Varley, J., Schullery, P. (Eds.), The Yellowstone Lake Crisis: Confronting a Lake Trout Invasion, A Report to the Director of the National Park Service, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, Wyoming, pp. 22 –27. Varley, J.D., Schullery, P., 1996. Freshwater wilderness. Yellowstone Laboratory and Museum Association. Vitousek, P., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of earth’s ecosystems. Science 277, 494 – 499. Williamson, M., 1996. Biological Invasions. Chapman and Hall, London, 244 p. Wuerthner, G., 1995. A fish story. National Parks, 25 – 27.
ANALYSIS
On the joint determination of biological and economic systems Chad Settle a,*, Thomas D. Crocker b, Jason F. Shogren b b
a Department of Economics, Uni6ersity of Tulsa, Tulsa, OK 74104 -3189, USA Department of Economics and Finance, Uni6ersity of Wyoming, Laramie, WY 82071 -3985, USA
Received 2 November 2001; received in revised form 12 April 2002; accepted 12 April 2002
Abstract Scarce natural resources and our choices to protect or develop them make ecological and economic systems jointly determined—human choices affect nature; nature affects human choices. This essay considers whether a dynamic model that integrates details of an economic system and an ecosystem with explicit feedback links between them yields significantly different results than does ignoring these links. We focus on the case of exotic invaders that put native species at risk in Yellowstone National Park. The results suggest that integration does matter— in each scenario, cutthroat trout populations differ in both magnitude and survival rates, depending on whether feedback is allowed between the two systems. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Economics; Ecology; Integration; Trout; Yellowstone; Exotic species
1. Introduction Despite the common rifts between the two disciplines, economics and the biological sciences are rife with similarities. Both are disciplines of limits — how to deal with scarcity. Whether it is a human reaction to a limited budget and unlimited wants or a fish’s response to limited food and unlimited appetite for reproduction, species must deal with their limits. The limiting factors in both disciplines drive their research efforts. Yet failure to account for joint influences upon these limits in * Corresponding author. Tel.: +1-918-631-3157; fax: + 1918-631-3546 E-mail address: [email protected] (C. Settle).
economic and biological systems can cause inaccurate perceptions of how each system works and provide misleading policy recommendations (Dasgupta et al., 2000). Joint determination creates a sequence of natural and human actions and reactions, and a feedback loop is born. The disturbances in one system set off repercussions in the other system, and these repercussions feedback into the system where the disturbances originated. The issue of risks to threatened and endangered species provides a vivid example. Conservation biologists often maintain that thresholds of species endangerments are functions of the present signs, trends and distributions of species’ populations and their likely interactions with habitats —strictly a bio-
0921-8009/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 0 0 9 ( 0 2 ) 0 0 1 0 5 - 2
302
C. Settle et al. / Ecological Economics 42 (2002) 301–311
logical question. We hold that this perspective is too narrow. Economic circumstances affect the quality of habitat. The circumstances that matter include the relative returns to human users from alternative sites, the relative returns from alternative uses on a particular site, and human welfare. Sites with low relative returns in their ‘highest and best’ use are more likely to be left undisturbed. Moreover, the rich can better afford to set aside quality habitat. Species survival is determined by economic as well as by biological parameters. Lacking unequivocal evidence to the contrary, the validity of separate treatments of the two parameter sets ought to be systematically demonstrated rather than routinely invoked. Models in which economic systems affect ecosystem outcomes are abundant in the economic literature. Fishery and forestry management models incorporate economic function into ecosystem representations (Clark, 1976). In the fishery context, these models often include predator –prey relations, selective harvesting or multiple species and may even introduce a spatial component which humans can influence. But these models are missing an important point of integration —feedbacks. Allowing for fishing pressure or harvest effort in a model accounts for how economic agents can alter the ecosystem. Having fishing pressure or harvesting effort as a constant, however, does not account for how humans adapt to a change in the fishery. With constant fishing effort, as fish populations fall due to an array of biological considerations, the harvest of fish also falls. Integrating economic systems and ecosystems via fishing effort captures this initial change. What it does not capture is how a change in one system can lead to a change in behavior in the other. When the fish species declines, will fishing effort actually be constant? When the fish population declines, an array of economic factors can cause humans to shift their efforts from one fish species to another or from fishing to other activities. This shift in behavior could lead to a different ecosystem steady state than if no account were taken of these feedbacks. This essay evaluates what taking account of the evolution of the details of jointly determined natural and human systems implies for a key, but
heretofore empirically untreated, question in the management of environmental and natural resources. Can a dynamic model that integrates an economic system and an ecosystem by formulating the details of feedback links between the two systems yield significantly different results than does the standard practice of giving short shrift to these links? We show how accounting for feedback between humans and nature affects the predicted ecological impacts— the population of a native prey species— caused by an exotic invader in Yellowstone Lake in Yellowstone National Park, Wyoming. The results show how acknowledgement of feedbacks might alter the core propositions, procedures, and public policy implications of both ecology and economics.
2. An application: exotic invaders in Yellowstone Lake Organisms that move beyond their traditional natural range can have undesirable ecological and economic consequences. Science has documented numerous examples of exotic plants and animals causing monetary and non-monetary damages (see Williamson, 1996). Consider a few classic examples. Field bindweed (Con6ol6ulus ar6ensis) is estimated to cause over $40 million in crop damages in Kansas every year (FICMNEW, 1998). The zebra mussel (Dreissena polymorpha) in the Great Lakes has significantly diminished phytoplankton biomass and harmed man-made structures (MacIsaac, 1996). The Nile perch (Lakes niloticus) has caused extinction of native fish and water quality problems in Lake Victoria. We now confront a similar problem in Yellowstone Lake, Wyoming, with the invasion of exotic lake trout (Sal6elinus namaycush). Yellowstone Lake is one of the last great inland fisheries in the United States for the native Yellowstone cutthroat trout (Oncorhynchus clarki bou6ieri ). Cutthroat trout are popular with fishermen and many predators such as ospreys (Pandion halieatus), white pelicans (Pelecanus erythrorhynchus), river otter (Lutra canadensis), and grizzly bears (Ursus arctos). In 1994, however, an angler caught a lake trout in Yellowstone Lake. Lake trout are an
C. Settle et al. / Ecological Economics 42 (2002) 301–311
exotic species to Yellowstone Lake— they prey upon but do not replace cutthroat trout in the food web. Fisheries biologists fear the lake trout population will expand and cause a severe decline in the cutthroat population (see Ruzycki and Beauchamp, 1997). If left unchecked, some biologists have predicted that this voracious exotic species could reduce the catchable-size cutthroat population from 2.5 million to 250,000– 500,000 within the near future (Kaeding et al., 1995). Furthermore, the ecosystem impact does not end with fewer cutthroat trout; grizzly bears, ospreys, eagles, river otters, and the other 40 species that rely on cutthroat as part of their food supply are put at risk (Schullery and Varley, 1995; Varley and Schullery, 1995, 1996). Most visitors to Yellowstone might not recognize the impact on cutthroats, but they are concerned with the loss of the larger, more charismatic species like grizzly bears that define Yellowstone for many people (Cherry and Shogren, 2001). This species interdependence could affect tourism in Yellowstone National Park. People come to Yellowstone to fish, bird watch, and they hope, see a grizzly bear. Some visitors might reduce the frequency of their visits, or they might enjoy the park less because of the changes within the ecosystem. An ecosystem externality may occur—the impact of lake trout on cutthroat trout shifts ecosystem equilibria, and the new equilibria affect the utility of park visitors through fewer cutthroats and the consequences for their predator species (see Crocker and Tschirhart, 1992). The species links within the ecosystem affect the utility of park visitors, whether they fish or not. Reducing risks to cutthroat trout, maintaining overall ecosystem integrity, and retaining visitor enjoyment provide ample justification for Yellowstone National Park managers to reduce pressure on the cutthroat trout population. By gillnetting lake trout, the park managers show, by maintaining a viable population of cutthroat trout or a large enough cutthroat trout population to sustain Yellowstone Lake as a cutthroat trout fishery, that they favor cutthroat trout over lake trout. What if accounting for the specifics of feedbacks between the two systems with and with-
303
out gillnetting leads to a difference in anticipated cutthroat trout populations? Then not addressing these feedbacks explicitly could lead Park managers to devote either too few or too many resources to the cutthroat trout problem. The question is whether accounting for the feedbacks between the Park’s ecological and economic systems provides additional useful data to managers who are considering incurring the costs of gillnetting the lake trout population in Yellowstone Lake.
3. Modeling biological –human interaction Traditionally, the specifics of threats to species and ecosystems have been estimated using one approach— the ‘damage function’ (DF) approach (see Freeman, 1993). The DF approach assumes the economic system and the ecosystem affect each other in a one-sided way. A change in the economic system is viewed as only changing the pressure on the ecosystem (Vitousek et al., 1997), or a change in the ecosystem is viewed as only changing the economic system (Daily, 1997). The DF approach, therefore, does not address the idea of co-evolution— the two-way interactions between human and natural systems (see Daly 1968; Norgaard, 1981).When an ecosystem changes, people change their behavior, which in turn reshapes the ecosystem, and so on. Ecosystem changes alter human productivity in the economic system. People recognize the change in their productivity when using the ecosystem and they adapt to this change, either by adapting the environment or by adapting to the environment. When people adapt, they alter the pressure they put on the ecosystem, leading to further changes in the ecosystem. The cycle continues. Can an explicit accounting of the specifics of these feedback links between the two systems yield different policy-relevant results than does assuming that no joint determination occurs? Though the broad outlines of ecological– economic system reciprocities have been acknowledged, the acknowledgements usually lump together key parameters of one or both systems. The lumping hinders predictions and evaluations
304
C. Settle et al. / Ecological Economics 42 (2002) 301–311
of which ecosystem services are to be usurped by humans, which waste flows are to be allowed to enter ecosystems, and which ecosystem biota and physical attributes are to be maintained. One can neither discriminate among the key parameters nor identify exactly how they enter the systems. Arguably, the contributions of natural environments to human well-being depend as much upon the complex details of ecological and economic states as upon the gross relationships between ecological and economic systems. Those analyses capable of admitting both the empirical detail and the jointness of ecological and economic systems are few (see Ayres and Kneese, 1969; Isard, 1972; Amir, 1979; Crocker and Tschirhart, 1992). These general equilibrium treatments work only with steady states. More recently a few authors (Swallow, 1996; Perrings, 1998; Sohngen and Mendelsohn, 1998; Brown and Layton, 2001) have set the structural details of jointness in a dynamic context. Here we describe a model and its results for two integrated systems: the economic system in Yellowstone National Park and the ecosystem in and around Yellowstone Lake. We have constructed a dynamic modeling framework incorporating many of the details of the flows inside each system and specifics of the reciprocal flows between the two systems. Incorporating the particulars of these links makes the model better reflect tradeoffs facing managers. We examine whether integrating economic systems and ecosystems matters, and if so, to what degree, by comparing the modeling results with and without accounting for the structure of feedbacks between the two systems.
4. Simulation model While exploiting a detailed analytical model of the Yellowstone Lake economic system and ecosystem, we use STELLA 2.0 software to simulate the importance of joint determination of the two species, lake and cutthroat trouts (Settle and Shogren (2002), provide the full mathematical specification of the model). Given the interactions and feedback loops between predator and prey
species and between species and humans, the full STELLA specification looks like a ball of yarn. We reduce the clutter for the reader by separating the full model into three main components: (1) lake trout; (2) cutthroat trout; and (3) human interactions. Components (1) and (2) capture the basic biological links between the species in and around Yellowstone Lake. Component (3) describes the human factors, including the interactions between humans and the ecosystem. We parameterized each component using population and consumption estimates for each species derived from the existing literature on the biology and socio-economics of Yellowstone Lake.1 These parameter estimates were used to build the interface between the economic system and the ecosystem. Now consider each component.
4.1. Model component 1: lake trout Diagram 1 shows the lake trout population and its direct biological links. The box labeled Lake Trout is a state variable indicating the population of lake trout in Yellowstone Lake at any given time. Deaths (LT natural deaths) are a flow out of the state variable, a decrease in the lake trout population. Births (LT births) are a flow into the state variable, an increase in population. The births are capped by a density dependence factor (DDFlt), the growth of lake trout is shown with a compensation model— the growth rate of lake trout declines with increases in the population of lake trout. The density dependence factor relates births of lake trout to its primary food source— cutthroat trout. As the cutthroat trout population declines, fewer lake trout can be supported in the lake. Now lake trout reallocate more time/energy to find food and less time/energy to spawn, which reduces spawning success. The reduction in spawning success reduces the number of births of lake trout. Note that the population of cutthroat trout also plays a key role in the density depen1 This research includes Blanchard (1980), Davenport (1974), Diem and Pugesek (1994), Kaeding (1995), Kaeding et al. (1995), Ruzycki and Beauchamp (1997), Swenson (1979), Stapp and Hayward (1998, 1999), Varley (1983), Varley and Schullery (1995, 1996), Wuerthner (1995).
C. Settle et al. / Ecological Economics 42 (2002) 301–311
dence factor, which we now consider in more detail.
305
the density-dependence factor limits the births of the species and places a carrying capacity on the cutthroat trout population. (2)
(1)
4.2. Model component 2: cutthroat trout
4.3. Model component 3: human impact
Diagram 2 shows the cutthroat trout population and its biology. The box labeled Cutthroat Trout is a state variable representing its population in the lake. Again, the two flows affecting population are births and natural deaths of cutthroat trout. A key difference between the cutthroat and lake trout components is that we have three species feeding upon cutthroat trout — lake trout, birds of prey, and grizzly bears. Therefore, the causes of cutthroat trout deaths include natural causes (CT natural deaths), predation by lake trout (CT killed LT), predation by birds of prey (CT killed B), and predation by grizzly bears (CT killed G). Each relationship is a function of the cutthroat trout population and the predatory species populations (state variables for these predatory species included in diagram 2). Lake trout are not likely to replace cutthroat trout in the diet of these predators since lake trout primarily stay in deep water and spawn in the lake instead of in the streams, where predators such as grizzly bears catch cutthroat trout (Stapp and Hayward, 1999). In addition, a density-dependence factor for cutthroat trout enters the model; once again the population of cutthroat trout is shown with a compensation model. Similar to the lake trout,
Now consider the human interaction with the ecosystem. Diagram 3 shows this interaction as captured by a representative visitor to Yellowstone National Park and its park manager. First, consider the visitor. Our visitor gains both direct and indirect benefits from species when visiting Yellowstone. The direct effects include: (1) fishing for lake trout (Human LT catch) and cutthroat trout (CT catch), which depends on the population of the fish and the time spent fishing (T F); 2 and (2) visiting the core attractions like Old Faithful and other geothermal activity, which we define as a park public good (X).The indirect effects include seeing birds of prey (Birds 6iewed) and grizzly bears (G 6iewed) while driving or fishing. The number of bird and grizzly bear sightings depends on species populations and time spent fishing (T F) and driving (T X).
2 Visitors are not divided into those who fish for lake trout and those who fish for cutthroat trout since Yellowstone Lake is a cutthroat trout fishery. People who intend to catch lake trout during a visit are able to go to Jackson Lake or other lake trout fisheries in northwestern Wyoming.
306
C. Settle et al. / Ecological Economics 42 (2002) 301–311
(3)
Our second human interaction is through the Park managers. These managers decide how to allocate a fixed budget for the park. In our model, they spend this fixed budget3 ($T) on two activities —improving the park public good4 ($X), and gillnetting the exotic lake trout to reduce pressure on the native cutthroat trout ($LT). The success of the gillnetting program, the number of lake trout killed by gillnetting (LT killed NPS), depends on both the budget spent gillnetting and on the lake trout population. This human interaction with the ecosystem allows us to capture the feedback links between the economic system and the ecosystem for Yellowstone Lake. These feedback links are captured by how the introduction of lake trout into Yellowstone Lake leads to changes in the ecosystem in and around Yellowstone Lake. The representative visitor responds to these changes by altering his behavior. When lake trout reduce the population of cutthroat trout, he finds it more difficult to 3 This fixed budget is always spent. The budgetary difference under the scenarios is the relative amount of the budget spent gillnetting lake trout versus improving the park public good. This fixed budget is also assumed constant. The policy effect on visitation is outside the scope of this model. 4 We use the condition of roads in the park as a proxy for the park public good.
catch cutthroat trout. Since he has less fishing success—the shadow price of fishing increases relative to the prices of alternative activities, he reallocates his time away from fishing at the Lake and toward other activities inside or away from the park where productivities have not declined, e.g. visiting Old Faithful. Therefore, the total number of fishing hours at Yellowstone Lake falls as the cutthroat population declines. This dynamic feedback loop does not jump from one steady state to another. The visitor reduces his time fishing in reaction to the changing ecosystem. Since the ecosystem is changing for several years, humans are altering their behavior for several years as well. Also, when humans alter their behavior, they change the pressure put on the fish species, which changes the path toward a new steady state for those species. This dynamic process continues until a steady state has been reached for both the ecosystem and the economic system— in which all species, including humans, are no longer reacting to a change in the ecosystem. As a comparative baseline, we presume that without such dynamic feedback loops the visitor does not respond to changes in the ecosystem. When lake trout reduce the population of cutthroat trout, assume he acts in the same manner as he did before lake trout were present. By measuring whether including feedbacks in our integrated model makes a difference in predicted fish populations, we can address the central question in this paper— how far off might the predictions be if we do not address the feedback loops between the two systems?
5. Results We consider three cases, each with and without feedbacks. First is the best-case scenario in which lake trout are immediately eliminated from Yellowstone Lake without cost. Cutthroat trout return to the world they had before lake trout were introduced. While infeasible in reality, the bestcase scenario defines the upper baseline on our indicators of well-being. Second, the worst-case scenario occurs when lake trout are left alone
C. Settle et al. / Ecological Economics 42 (2002) 301–311
307
Fig. 1.
without any interference from the National Park Service. Park managers do not attempt to help cutthroat trout survive. Instead, lake trout and cutthroat trout are left to reach their own steady state equilibrium. This worst-case produces the lower benchmark since it produces the smallest cutthroat trout population, and the lowest odds of cutthroat trout survival. Third, our policy scenario has the National Park Service reducing the risk to cutthroat trout by gillnetting lake trout.5 Assume the Service’s current level of expenditures on gillnetting lake trout is continuous and perpetual. We use the population of cutthroat trout as a yardstick.6 If the population of cutthroat trout is similar with and without feedbacks, an accounting of any feedbacks between the Yellowstone Lake ecosystem and economic system might not be worth the effort. If the resulting populations differ significantly within all cases, the accounting is more likely to provide the foundation for an improved understanding of the behavior of both systems. Table 1 and Fig. 1 summarize the results for each scenario. Under the best-case scenario without feedbacks, the steady state population of cutthroat trout is about 2.7 million. With feedbacks, 5 Gillnetting is used as the policy alternative since it is the method used by the National Park Service, and research determining the costs of reducing the lake trout population under various alternatives found gillnetting to be the lowest budgetary cost method (Varley and Schullery, 1995). 6 Listed populations are the steady state populations. The time to approach the steady state populations varies by scenario, with most approaching the steady state within 100 –200 years.
Table 1 Resulting cutthroat trout populations with and without feedbacks
Best-case scenario Policy scenario Worst-case scenario
No feedbacks
Feedbacks
2,700,000 1,900,000 900,000
3,400,000 2,300,000 0
the steady state population is about 3.4 million. The difference arises from fishermen’s behaviors. Without feedback, fishermen continue to fish as before, putting constant pressure on the cutthroat. With feedback, fishermen exploiting declining cutthroat populations adapt by fishing less and visiting other attractions more. Reduced human pressure on the cutthroat allows its population to increase by an amount greater than with constant fishing pressure. The resulting population of cutthroat trout is therefore greater with feedback from human adaptation. The results are similar for the policy scenario. Without feedback, visitors spend a constant amount of time fishing and put a constant pressure on cutthroat trout. The steady state population of cutthroat trout is 1.9 million. With feedback, however, the cutthroat population increases to 2.3 million. Again accounting for feedback matters—predicted populations are 20% greater. When the cutthroat trout population declines, fishermen are less successful. They respond by fishing less for cutthroat trout, which in turn allows the cutthroat trout population to increase over what it would be under constant fishing pressure.
C. Settle et al. / Ecological Economics 42 (2002) 301–311
308
Table 2 How system shocks affect the steady state populations of cutthroat trout Visitation at 5 million
Best-case scenario Policy scenario Worst-case scenario
No feedbacks Feedbacks
No fishing
2,300,000
3,100,000
4,700,000
1,700,000
2,000,000
2,700,000
1,100,000
0
0
The worst-case scenario tells a similar story but with a policy conclusion opposite that for the best-case and the policy scenarios. Assuming the lake trout population increases unabated, the cutthroat trout population falls dramatically. Without feedback, visitors put constant pressure on the cutthroat trout, and its steady state population falls to 0.9 million. With feedback, visitors again reduce the pressure on cutthroat trout, which should increase its steady state population. But a powerful countervailing effect exists—when people fish less, the incidental catch of lake trout also declines. As in the best-case and the policy scenarios, when visitors fish less, the incidental catch of lake trout drops, which in turn allows the larger population of lake trout to kill more cutthroat trout. But in this worst-case, the effects of less fishing pressure on the lake trout population dominates its cutthroat effect, supporting Volterra’s principle (Braun, 1975). Now the steady state population of cutthroat trout drops to zero. What matters is that when people shift their time away from fishing as the cutthroat trout population declines and the lake trout population increases, the incidental catch for fishermen of lake trout becomes an important control of the lake trout population. People shift away from fishing and that important control of lake trout declines. While one usually thinks of fishing pressure to be a pressure on cutthroat trout, in this case the more critical pressure is on lake trout, which were a higher proportion of fishermen’s catch. A no-feedback model suggests a healthier outcome than might actually exist—nearly 1 mil-
lion cutthroat trout versus none is a significant difference however one measures it. Feedbacks yield both different magnitudes and survival rates, suggesting that one discounts the importance of feedback loops at one’s own risk. We see that the system yields different outcomes under the scenarios with and without feedbacks—now consider the robustness of this result to two external shocks to the environment: (1) more fishing pressure due to increased visitation to the park; and (2) no fishing pressure due to a complete ban on fishing. First, we implement the visitation shock by increasing visitation to the park from 3 to 5 million visitors, the estimated visitors in 50 years if current trends continue (2050).7 Table 2 shows that the same patterns emerge with the population shock. More visitation increases fishing pressure, which lowers steady state populations for cutthroat trout, but we still have a positive population of cutthroat trout in every scenario except the worst-case scenario with feedbacks. Under the best-case scenario, the steady state population of cutthroat trout is 2.3 and 3.1 millions without and with feedbacks. With the policy scenario, the steady state population of cutthroat trout is 1.7 and 2 millions without and with feedbacks. Finally, under the worst-case scenario, the steady state population of cutthroat trout is 1.1 million and 0 without and with feedbacks. Shocking the system by adding a lot more visitors does not alter the direction of the results, only the size of the steady state populations. Feedbacks still matter. The second shock to the system is to ban fishing at the Lake— visitors are prohibited from catching either lake trout or cutthroat trout. Without fishing pressure on the ecosystem, the 7 We estimated the anticipated level of visitation in 2050 by running a univariate time series model on National Park Service visitation data between 1970 and 2000, and then extrapolated the results to 2050. The data indicated the expected level of visitation at 2050 was between 4 and 6 million, depending on which date was used as the starting value. Estimated visitation based on data from 1970 to 2000 was about 5 million, from 1980 to 2000, visitation was about 6 million; and from 1990 to 2000, anticipated visitation was about 4 million. We use the midpoint of these three estimates, i.e. 5 million. Data was collected from Yellowstone National Park’s webpage (www.nps.gov/yell).
C. Settle et al. / Ecological Economics 42 (2002) 301–311
steady state population of cutthroat trout increases under both the best-case scenario and the policy scenario to 4.7 and 2.7 millions.8 Under the worst-case scenario, the steady state population of cutthroat trout falls to 0, once again highlighting the importance of the inadvertent catch by visitors of the predatory lake trout. Neither visitation nor fishing ban shocks reduce the importance of feedbacks in modeling economic system–ecosystem interactions.
6. Concluding remarks Mark Twain said ‘‘Pain there was none, nor infirmity, nor any physical signs to mark the flight of time … in Eden’’. But when ‘‘Eve reached for an apple!— oh farewell Eden, and your sinless joys, come poverty and pain, hunger and cold heartbreak, … strife, …, age, weariness, remorse …’’ (Twain, 1991). Our aspirations and our temptations have cursed us with scarcity. We yield to human arrogance and ignorance— reaching for and biting the apple, and in so doing, we affect nature, ‘oh farewell Eden’, and thus cause nature to affect us—‘come poverty and pain’. This curse of Adam and Eve says that scarcity and the choices it necessitates make ecological and economic systems jointly determined: human choices affect nature; nature affects human choices. The curse of scarcity matters in the management of exotic invaders in Yellowstone Lake. Our results suggest that a dynamic model that integrates biological and economic systems with explicit treatment of some of the complexities of feedback links between the two systems yields significantly different results than when one ignores these links. In each scenario, cutthroat trout populations differ in both magnitude and survival rates once feedback is allowed between the two systems. For the best-case and the policy scenarios, no feedback leads to a lower estimate of the steady state population of cutthroat trout than 8 Eliminating fishing as an option yields no differences between feedbacks and no feedbacks since the option to fish is gone— all visitors spend no time fishing once it has been prohibited.
309
does feedback. But with the worst-case scenario, ignoring feedback actually leads the model to a higher estimate of the cutthroat population than does the model with feedback— a steady state population of 1 million versus none. This suggests that basing policy recommendations in Yellowstone Lake on data from models without feedback puts cutthroats at greater risk than would be true if feedback was explicitly considered. The potential policy relevance of the joint determination of biological systems and economic systems and the feedback it implies extends well beyond the Yellowstone Lake case. Policy makers should take the curse of Adam and Eve seriously. Consider AIDS, for example. Traditional epidemiological practice treats a person’s decisions concerning frequency of sexual contact and number of partners to be independent of the prevalence of the disease. But concede that people make choices about contacts and partners based on the odds of infection and the benefits and costs of what they can do to influence these odds. The predicted incidence of AIDS will be misestimated if the determinants of people’s choices are not taken into account. The light cast by making joint determination visible raises features of the AIDS system which would otherwise go unnoticed. Routine dismissal of the joint determination could compel similar errors in valuing alternative environmental states. For example, the gains from reducing the risk of lead poisoning in U.S. children have been estimated to at least double when one accounts for parents’ decisions to reduce children’s exposures to and body burdens of lead (Agee and Crocker, 1996). The inclusion of the behavioral interactions of economics and ecosystems for protection of North Carolina brackish wetland acreages leads to an average ninefold increase in calculated benefits (Swallow, 1996). In each case, the added benefits result from combinations of protection expenditures and natural asset price increases avoided, and the effects upon natural asset productivity of human adaptations to these avoided expenditures and price increases. The routine acceptance of the separability of natural events and human adaptations is not just one move in the long line of trade-offs between tractability and completeness which effective sci-
310
C. Settle et al. / Ecological Economics 42 (2002) 301–311
ence in support of policy requires. The separability question runs much deeper. The habitual invocation of separability obstructs economists’ and other social scientists’ opportunity — and obligation — to help define the thresholds of human health and ecosystem resilience, which underpin environmental policy. Like the snake in the apple tree of Eden, the invocation does a serious disservice to the common weal. One has to go beyond recognizing that the snake exists—actively seeking out more ways to map the details of scarcity and the jointness of natural and human systems is one path. In their extant studies of environmental –economy interactions, the two disciplines are biased in opposite directions. Ecologists are prone to regard anthropogenic environmental impacts as exogenous parameter shocks; they disregard the complexity of economic systems. Economists tend to view the environment as an aggregate or macro form in which interactions with the form can be dismissed; they therefore disregard the complexity of the ecological system. Explicit recognition of joint determination has an additional advantage: both ecology and economics can be better motivated to account for the complexities of both ecosystems and economic systems.
Acknowledgements We thank Greg Hayward, Paul Stapp, Wayne Hubert, the Yellowstone National Park for cooperation, and the USEPA EPSCOR Grant c R826281-09-0 and Stroock Professorship for financial support.
References Agee, M., Crocker, T., 1996. Parental altruism and child lead exposure. Journal of Human Resources 31, 677 –691. Amir, S., 1979. Economic interpretation of equilibrium concepts in ecological systems. Journal of Social and Biological Structures 2, 293 –314. Ayres, R., Kneese, A., 1969. Production, consumption, and externalities. American Economic Review 59, 282 –297. Blanchard, B.M., 1980. Grizzly bear – habitat relationships in the Yellowstone area. International Conference of Bear Reservation and Management 5, 118 –123.
Braun, M., 1975. Differential Equations and Their Applications: An Introduction to Applied Mathematics. SpringerVerlag, New York, 718 p. Brown, G., Layton, D.F., 2001. A market solution for preserving biodiversity: the Black Rhino. In: Shogren, J.F., Tschirhart, J. (Eds.), Protecting Endangered Species in the United States: Biological Needs, Political Realities, Economic Choices. Cambridge University Press, New York, NY, pp. 32– 50. Cherry, T.L., Shogren, J.F., 2001. Invasive species management for the Yellowstone Lake ecosystem: what do visitors think? Yellowstone Science 9, 11 – 15. Clark, C., 1976. Mathematical Bioeconomics: the Optimal Management of Renewable Resources. Wiley, New York, 352 p. Crocker, T.D., Tschirhart, J., 1992. Ecosystems, externalities, and economies. Environmental and Resource Economics 2, 551 – 567. Daily, G. (Ed.), 1997. Nature’s Services: Societal Dependence on Natural Ecosystems. Island Press, Washington DC, 392 p. Daly, H., 1968. On economics as a life science. Journal of Political Economy 76, 392 – 406. Dasgupta, P., Levin, S., Subchenco, J., 2002. Economic pathways to sustainability. Bioscience 50, 339 – 345. Davenport, M.B., 1974. Piscivorous Avifauna on Yellowstone Lake, Yellowstone National Park. Completion report to Yellowstone National Park, Yellowstone Park, Wyoming, USA. Diem, K.L., Pugesek, B.H., 1994. American white pelicans at the Molly Islands, in Yellowstone National Park: twentytwo years of boom-and-bust breeding, 1966 – 1987. Colonial Waterbirds 17 (2), 130 – 145. Federal Interagency Committee for the Management of Noxious and Exotic Weeds (FICMNEW), 1998. Invasive plants: changing the landscape of America, Fact Book. FICMNEW, Washington DC. Freeman III, A.M., 1993. The Measurement of Environmental and Resource Values: Theory and Methods, Resources for the Future, Washington DC. Isard, W., 1972. Ecologic – Economic Analysis for Regional Development. The Free Press, New York, NY, 270 p. Kaeding, L.R., 1995. Annual Project Report for 1994. Fishery and Aquatic Management Program, Yellowstone National Park. Kaeding, L., Boltz, G., Carty, D., 1995. Lake trout discovered in Yellowstone Lake, The Yellowstone Lake crisis: confronting a lake trout invasion. In: Varley, J., Schullery, P. (Eds.), A Report to the Director of the National Park Service, Yellowstone Center for Resources National Park Service, Yellowstone National Park, Wyoming, pp. 4 – 11. MacIsaac, H., 1996. Potential abiotic and biotic impacts of zebra mussels on the Inland Waters of North America. Amercian Zoology 36, 287 – 299. Norgaard, R.B., 1981. Sociosystem and ecosystem coevolution in the Amazon. Journal of Environmental Economics and Management 8, 238 – 254.
C. Settle et al. / Ecological Economics 42 (2002) 301–311 Perrings, C., 1998. Resilience in the dynamics of economy – environment systems. Environmental and Resource Economics 11, 503 – 520. Ruzycki, J., Beauchamp, D., 1997. A bioenergetics modeling assessment of the lake trout impact in Yellowstone Lake. In: Hamre, R. (Ed.), Wild Trout IV, Trout Unlimited and Federation of Fly Fishers. Fort Collins, CO, pp. 127 – 133. Schullery, P., Varley, J., 1995. Cutthroat trout and the Yellowstone Lake ecosystem. In: Varley, J., Schullery, P. (Eds.), The Yellowstone Lake Crisis: Confronting a Lake Trout Invasion, A Report to the Director of the National Park Service, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, Wyoming, pp. 12 –21. Settle, C., Shogren, J., 2002. Accounting for Feedback between Biological and Economic Systems: The Case of Exotic-Native Conflicts in Yellowstone Lake. Working paper, University of Tulsa. Sohngen, B., Mendelsohn, R., 1998. Valuing the impact of large-scale change in a market: the effect of climate change on U.S. timber. American Economic Review 88, 686 –710. Stapp, P., Hayward, G.D., 1998. Estimates of Predator Consumption of Yellowstone Cutthroat Trout. Working paper, Department of Zoology, University of Wyoming. Stapp, P., Hayward, G.D., 1999. Effects of Introduced Lake Trout in Yellowstone Lake: Insights from Demographic Analysis of Yellowstone Cutthroat Trout. Working paper, Department of Zoology, University of Wyoming. Swallow, S., 1996. Resource capital theory and ecosystem economics: developing nonrenewable habitats with hetero-
311
geneous quality. Southern Economic Journal 63, 106 – 123. Swenson, J.E., 1979. Factors affecting status and reproduction of ospreys in Yellowstone National Park. Journal of Wildlife Management 43 (3), 595 – 601. Twain, M., 1991. The Bible according to Mark Twain. In: Baetzhold, H.Y., McCollough, J.B. (Eds.), Writings on Heaven, Eden, and the Flood, The University of Georgia Press, Athens, GA, pp. 53, 66. Varley, J.D., 1983. The use of restriction regulations in managing wild salmonoids in Yellowstone National Park, with particular reference to cutthroat trout. In: Proceedings of the Olympic Wild Fish Conference, Penadweuh College, Port Angeles, Washington, pp. 145 – 156. Varley, J., Schullery, P., 1995. Socioeconomic values associated with the Yellowstone Lake cutthroat troat. In: Varley, J., Schullery, P. (Eds.), The Yellowstone Lake Crisis: Confronting a Lake Trout Invasion, A Report to the Director of the National Park Service, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, Wyoming, pp. 22 –27. Varley, J.D., Schullery, P., 1996. Freshwater wilderness. Yellowstone Laboratory and Museum Association. Vitousek, P., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of earth’s ecosystems. Science 277, 494 – 499. Williamson, M., 1996. Biological Invasions. Chapman and Hall, London, 244 p. Wuerthner, G., 1995. A fish story. National Parks, 25 – 27.