SYSTEM DYNAMICS
SD is a lot about understanding how things work differently compared to what expected.
More road do not always lead to less congestion? In the short run this might lead to less congestion, but in
the long run it can lead to more people using private transportation.
Antilock brakes do not always result in less severe accidents? People started to trust these systems and
they started to brake latter and keep less distance between cars, so the overall effect is not improved by
this new system.
Things that work differently than expected
There is always an unintended side effect.
It is not enough to look at some restricted areas (like the belly of the elephant and not the whole elephant).
Rather than adopting a narrow, linear point of view, we need to look at the bigger picture. To look at the
whole system, and anticipate effects that might undermine our proposed solutions. This is what system
dynamics is all about.
1. Everything is connected; hence what happens in one part of the system affects another. For example,
evaluation of a school improvement initiative may have to examine not just what takes place inside the
school system, but also what happens in the broader community outside.
2. Cause and effect is not a linear, one-directional process; it is much more iterative. Does improving
health of families improve their economic productivity or the other way around? The answer is probably
both. We are increasingly learning that honing in on causation and attempting attribution is a herculean
task, as there aren’t clear and straightforward links.
3. Context matters; a lot! What used to be considered “noise” in models of social change is now recognized
as a core factor that can make or break an intervention. Teachers in schools with supportive conditions, for
example, are shown to be performing at higher levels than those (even highly qualified ones) in other non-
supportive schools
Why do things work out differently than expected?
• Dynamic complexity (e.g., information delays)
• Mental models
,MISMATCH OF CHARACTERISTICS
POLICY RESISTANCE when things work out differently because there are also side effects.
Delays caused by dynamic complexity dramatically reduce performance.
Mental models we use are frequently characterized by:
• Short time horizon
• Narrow boundaries
• Open loop perspective
• Mono-causal explanation
• Poor understanding of chance & uncertainty, time delays, stock and flows, non-linearities
Dynamic—real world—systems are:
• Dynamic (incl. significant delays)
• Tightly coupled
• Governed by feedback
• Nonlinear
• Self-Organizing
• Adaptive
• Evolving
Short term horizon we like sentences like: ‘we do thinks like that’…
System dynamics offers a set of tools, directed toward overcoming this ‘issues’:
Systems thinking - to solve dynamically complex problems - by means of:
o Causal loop diagramming
o System dynamics simulation modeling
Causal loop diagramming when things get to complex we can move to system dynamics simulation
modeling
Linear view of the problem, there is a situation what we expect is not like things are
,The second one is more complex than the other one.
• Many of the problems we now face arise as unanticipated side effects of our own past actions…
• The world is complex, governed by feedback, delays and non-linearities
• To anticipate such side-effects, we need to look at the bigger picture!
This is what systems thinking is about!
Ripple effect of motor clubs they just moved across the border they left and they settled in another
place this is an unattended side effect
To disclose possible side-effects we need to look at the bigger picture...
1. Everything is connected; hence what happens in one part of the system affects another. For example,
evaluation of a school improvement initiative may have to examine not just what takes place inside the
school system, but also what happens in the broader community outside.
2. Cause and effect is not a linear, one-directional process; it is much more iterative. Does improving
health of families improve their economic productivity or the other way around? The answer is probably
both. We are increasingly learning that honing in on causation and attempting attribution is a herculean
task, as there aren’t clear and straightforward links.
3. Context matters; a lot! What used to be considered “noise” in models of social change is now recognized
as a core factor that can make or break an intervention. Teachers in schools with supportive conditions, for
example, are shown to be performing at higher levels than those (even highly qualified ones) in other non-
supportive schools.
Thinks can work out differently then expected.
The rental house example
• Goal of Dutch government (around 08/09): every citizen must
be able to afford a place to live
• Decision: keep rents low
• Goals of Corporations (that build and rent out houses):
continuity/making profit
• Decision of Corporations: stop building new houses because of
low or no profit
• New Situation: scarcity of rental houses, people with low
incomes cannot afford a place to live
, This example shows what is meant by "policy resistance":
o The tendency for interventions to be delayed, diluted, or defeated by the response of
the system to the intervention itself.
Policy resistance arises because we often do not understand the full range of feedbacks operating in the
system (over time incl. delays)
System Dynamics (systems thinking) is about creating an understanding of the workings of the dynamics
within a system
Dynamics of a System
Dynamics of a system are determined by:
• Feedback processes
• (Stock and flow) Structures
• Time delays
• Nonlinearities
There are only two types of feedback all behavior arises from two types of feedback loops
The last one can also decline in an accelerating way!
Reinforcing loops are self-reinforcing, they can spiral over time
These loops are also known as positive feedback loops
”Positive” because reinforcing loops amplify change
Balancing loops are self-stabilizing
These loops are also known as negative feedback loops
”Negative” because balancing loops counteract change
SD is a lot about understanding how things work differently compared to what expected.
More road do not always lead to less congestion? In the short run this might lead to less congestion, but in
the long run it can lead to more people using private transportation.
Antilock brakes do not always result in less severe accidents? People started to trust these systems and
they started to brake latter and keep less distance between cars, so the overall effect is not improved by
this new system.
Things that work differently than expected
There is always an unintended side effect.
It is not enough to look at some restricted areas (like the belly of the elephant and not the whole elephant).
Rather than adopting a narrow, linear point of view, we need to look at the bigger picture. To look at the
whole system, and anticipate effects that might undermine our proposed solutions. This is what system
dynamics is all about.
1. Everything is connected; hence what happens in one part of the system affects another. For example,
evaluation of a school improvement initiative may have to examine not just what takes place inside the
school system, but also what happens in the broader community outside.
2. Cause and effect is not a linear, one-directional process; it is much more iterative. Does improving
health of families improve their economic productivity or the other way around? The answer is probably
both. We are increasingly learning that honing in on causation and attempting attribution is a herculean
task, as there aren’t clear and straightforward links.
3. Context matters; a lot! What used to be considered “noise” in models of social change is now recognized
as a core factor that can make or break an intervention. Teachers in schools with supportive conditions, for
example, are shown to be performing at higher levels than those (even highly qualified ones) in other non-
supportive schools
Why do things work out differently than expected?
• Dynamic complexity (e.g., information delays)
• Mental models
,MISMATCH OF CHARACTERISTICS
POLICY RESISTANCE when things work out differently because there are also side effects.
Delays caused by dynamic complexity dramatically reduce performance.
Mental models we use are frequently characterized by:
• Short time horizon
• Narrow boundaries
• Open loop perspective
• Mono-causal explanation
• Poor understanding of chance & uncertainty, time delays, stock and flows, non-linearities
Dynamic—real world—systems are:
• Dynamic (incl. significant delays)
• Tightly coupled
• Governed by feedback
• Nonlinear
• Self-Organizing
• Adaptive
• Evolving
Short term horizon we like sentences like: ‘we do thinks like that’…
System dynamics offers a set of tools, directed toward overcoming this ‘issues’:
Systems thinking - to solve dynamically complex problems - by means of:
o Causal loop diagramming
o System dynamics simulation modeling
Causal loop diagramming when things get to complex we can move to system dynamics simulation
modeling
Linear view of the problem, there is a situation what we expect is not like things are
,The second one is more complex than the other one.
• Many of the problems we now face arise as unanticipated side effects of our own past actions…
• The world is complex, governed by feedback, delays and non-linearities
• To anticipate such side-effects, we need to look at the bigger picture!
This is what systems thinking is about!
Ripple effect of motor clubs they just moved across the border they left and they settled in another
place this is an unattended side effect
To disclose possible side-effects we need to look at the bigger picture...
1. Everything is connected; hence what happens in one part of the system affects another. For example,
evaluation of a school improvement initiative may have to examine not just what takes place inside the
school system, but also what happens in the broader community outside.
2. Cause and effect is not a linear, one-directional process; it is much more iterative. Does improving
health of families improve their economic productivity or the other way around? The answer is probably
both. We are increasingly learning that honing in on causation and attempting attribution is a herculean
task, as there aren’t clear and straightforward links.
3. Context matters; a lot! What used to be considered “noise” in models of social change is now recognized
as a core factor that can make or break an intervention. Teachers in schools with supportive conditions, for
example, are shown to be performing at higher levels than those (even highly qualified ones) in other non-
supportive schools.
Thinks can work out differently then expected.
The rental house example
• Goal of Dutch government (around 08/09): every citizen must
be able to afford a place to live
• Decision: keep rents low
• Goals of Corporations (that build and rent out houses):
continuity/making profit
• Decision of Corporations: stop building new houses because of
low or no profit
• New Situation: scarcity of rental houses, people with low
incomes cannot afford a place to live
, This example shows what is meant by "policy resistance":
o The tendency for interventions to be delayed, diluted, or defeated by the response of
the system to the intervention itself.
Policy resistance arises because we often do not understand the full range of feedbacks operating in the
system (over time incl. delays)
System Dynamics (systems thinking) is about creating an understanding of the workings of the dynamics
within a system
Dynamics of a System
Dynamics of a system are determined by:
• Feedback processes
• (Stock and flow) Structures
• Time delays
• Nonlinearities
There are only two types of feedback all behavior arises from two types of feedback loops
The last one can also decline in an accelerating way!
Reinforcing loops are self-reinforcing, they can spiral over time
These loops are also known as positive feedback loops
”Positive” because reinforcing loops amplify change
Balancing loops are self-stabilizing
These loops are also known as negative feedback loops
”Negative” because balancing loops counteract change
