104
limited in the variety and complexity of individual-level heterogeneity and interactions
that can be considered.
2
In addition to being constrained in terms of the amount of heterogeneity that can
be considered, there are other potential limitations of simplifying human decisionmaking. Optimization assumes that agents are maximizing or minimizing a known
objective, which may not reflect the underlying behavior. Assumptions of optimizing
behavior can be made more realistic by imposing constraints, e.g., by accounting for
uncertainty and then assuming a process by which individuals form expectations over
future unknown conditions or states of the world. However, this expectation formation
process is often assumed to be highly simplified and may not account for key processes, such as learning over time or through interactions with others.
Furthermore, while profit and other economic factors often play an important
role in individuals’ decision-making processes, people consider other factors in
their decision-making process as well. For example, farmers may use their land in
such a way that takes the environmental consequences of their decisions into
account, even if their land use strategy is not the one that maximizes profit. Similarly,
people may choose gasoline-inefficient vehicles over smaller, more efficient vehicles to signal status or other aspects of their identities, even when inefficient vehicles are economically irrational choices.
As discussed in Sect. 1.3.1, the FEW system domain is subject to policy intervention at all levels of government. Formal models of these coupled human and
natural systems often inform policy recommendations for FEW system management. As a result, relying on inaccurate representations of human behavior may lead
to policies that fail to achieve their stated aims or policies that are not implemented
at all. For example, economic models of policies intended to mitigate the effects of
climate change often present the economic costs of strategies to address climate
change without accounting for the potential benefits of new technologies, such as
economic growth or increases in employment. Additionally, models designed to
inform global climate change policies have traditionally assumed a market discount
rate (see Sect. 5.3.3), the rate at which the market discounts future economic returns
(typically the prevailing interest rate), rather than a social discount rate, which
would reflect individuals’ sense of moral obligation and concern for future generations. The use of a market discount rate underweights future environmental costs
and can have substantially different implications for policy.
Furthermore, studies that predict policy impacts often do not consider distributional effects, such as whether and how the policy change may be felt by high-versus
low-income populations, or how people in urban centers and rural areas may be
differentially impacted (See Sects. 5.3.2 and 18.4 for case studies that illustrate
distributive effects in FEW policy at the city-scale).
Policies that have different distributional impacts may result in conflict as groups
compete for increasingly scarce resources such as water or land, triggering human
2 See Irwin and Wrenn (2014) for a discussion of equilibrium-based and other modeling approaches,
including agent-based models, in the context of land use decision-making and land change
systems.
M. Doidge et al.
limited in the variety and complexity of individual-level heterogeneity and interactions
that can be considered.
2
In addition to being constrained in terms of the amount of heterogeneity that can
be considered, there are other potential limitations of simplifying human decisionmaking. Optimization assumes that agents are maximizing or minimizing a known
objective, which may not reflect the underlying behavior. Assumptions of optimizing
behavior can be made more realistic by imposing constraints, e.g., by accounting for
uncertainty and then assuming a process by which individuals form expectations over
future unknown conditions or states of the world. However, this expectation formation
process is often assumed to be highly simplified and may not account for key processes, such as learning over time or through interactions with others.
Furthermore, while profit and other economic factors often play an important
role in individuals’ decision-making processes, people consider other factors in
their decision-making process as well. For example, farmers may use their land in
such a way that takes the environmental consequences of their decisions into
account, even if their land use strategy is not the one that maximizes profit. Similarly,
people may choose gasoline-inefficient vehicles over smaller, more efficient vehicles to signal status or other aspects of their identities, even when inefficient vehicles are economically irrational choices.
As discussed in Sect. 1.3.1, the FEW system domain is subject to policy intervention at all levels of government. Formal models of these coupled human and
natural systems often inform policy recommendations for FEW system management. As a result, relying on inaccurate representations of human behavior may lead
to policies that fail to achieve their stated aims or policies that are not implemented
at all. For example, economic models of policies intended to mitigate the effects of
climate change often present the economic costs of strategies to address climate
change without accounting for the potential benefits of new technologies, such as
economic growth or increases in employment. Additionally, models designed to
inform global climate change policies have traditionally assumed a market discount
rate (see Sect. 5.3.3), the rate at which the market discounts future economic returns
(typically the prevailing interest rate), rather than a social discount rate, which
would reflect individuals’ sense of moral obligation and concern for future generations. The use of a market discount rate underweights future environmental costs
and can have substantially different implications for policy.
Furthermore, studies that predict policy impacts often do not consider distributional effects, such as whether and how the policy change may be felt by high-versus
low-income populations, or how people in urban centers and rural areas may be
differentially impacted (See Sects. 5.3.2 and 18.4 for case studies that illustrate
distributive effects in FEW policy at the city-scale).
Policies that have different distributional impacts may result in conflict as groups
compete for increasingly scarce resources such as water or land, triggering human
2 See Irwin and Wrenn (2014) for a discussion of equilibrium-based and other modeling approaches,
including agent-based models, in the context of land use decision-making and land change
systems.
M. Doidge et al.