QUERY Definition of root cause

This forum contains all archives from the SD Mailing list (go to http://www.systemdynamics.org/forum/ for more information). This is here as a read-only resource, please post any SD related questions to the SD Discussion forum.
Jack Harich <register@thwink.
Member
Posts: 39
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jack Harich <register@thwink. »

Posted by Jack Harich <register@thwink.org>

Jean-Jacques Laublé wrote:
> Hi Jack
> You write :
> <This is because the field still clings to phrases
> <like ""dynamic hypothesis"" of what is causing the problem instead of root
> <cause and feels that if a model can reproduce the symptoms, then it must
> <contain the root cause.
>
> I do not think that anybody pretends that reproducing the symptoms does
> guarantee anything.

Perhaps you came in late on this thread. One reply, from Bill Braun, was:

""Jack Harich asks about the definition of root cause. Broadly stated, it
would be the structure that produces the reference mode up to the
present and, if policies were left as is, would result in the ""feared""
mode in the future.""

Another reply, from Jack Homer, said: ""We call 'root cause' the dynamic
hypothesis.""

Also, Sterman writes that a key SD process step is: ""Formulate a dynamic
hypothesis that explains the dynamics as endogenous consequences of the
feedback structure.""


> The notion of dynamic hypothesis is not part of the SD paradigm but
> belongs
> to the most known and taught method of building models.

Thanks. I think the concept of a dynamic hypothesis has been central to
SD from the start.

Without using the term ""dynamic hypothesis"" Forrester wrote in Urban
Dynamics, 1969, page 113 that: ""The first step in modeling is to
generate a model that creates the problem. Only if we understand the
processes leading to the difficulties can we hope to restructure the
system so that the internal processes lead in a different direction. If
the model is to create the difficulties, it must contain all the
interacting relationships necessary to lead the system into trouble. The
troubles are not imposed on the system from outside the structure being
modeled. The model will be a closed model which is not dependent for its
inherent characteristic behavior on any variables transmitted across its
boundary from the external world.""

This says the model contains ""all the interacting relationships
necessary to lead the system into trouble"" and ""The troubles are not
imposed on the system from outside."" In other words, the model contains
the underlying causes, which I'm calling the root causes. If the
troubles are not imposed from the outside, then it follows that what
imposes the troubles must be inside the model.

Unless I'm missing something, ""The first step in modeling is to generate
a model that creates the problem"" seems to be the same as ""The first
step in modeling is to generate a model that reflects your dynamic
hypothesis of what it is that is causing the problem.""

Ergo, it follows that a dynamics hypothesis model is assumed to contain
the root cause. My point is that particularly in difficult problems, it
may not. Ability of a model to reproduce a system's behavior does not
guarantee it contains the root causes. But far too often, modelers and
model users assume otherwise.

This is a contentious point, so here's an example: Consider a model with
a constant, such as the infectivity rate for a virus. If the problem
deals with spread of a disease, and the virus is the disease, then the
infectivity rate is one of several root causes. The solution may center
on how to lower human infection rates to the point where an epidemic
cannot occur. This would resolve the root cause by reducing the power of
the virus to cause an epidemic. So in this example, the model appears to
contain the root cause, and easily mimics system behavior in epidemics.

But along comes a creative thinker. She asks ""Why is the infectivity
rate for that virus so high? That's a 20th century virus. It appeared
due to favorable evolutionary circumstances."" Subsequent investigation
shows that standard treatments were not knocking out the original virus
fast enough. This gave it time to mutate in patients. This led to new
more virulent strains. One of these was the one in the model. Here the
root cause is the standard treatments. They were insufficient to prevent
mutation. A second model could be build to include this. Or you could go
deeper in the root cause analysis, and ask ""Why did the social system
allow this to happen?""

The point is that the first model reproduced the problem's symptoms, but
did not contain the ""true"" root cause.

One might quibble that they didn't take the starting time of the model
back far enough. If they did, they would have had to include treatment
and mutation, and the model would have contained the true root cause.
But if the problem being modeled was an epidemic, there is no apparent
need to start the model long ago, when the virus was born. There is only
the need to model the current epidemic under investigation.

Every constant, every equation in a model is a candidate for probing
deeper. This can possibly lead to a bonafide root cause.

As another example, in Forrester's Urban Decay model, a reasonable root
cause would have been related to the question: Why did America's urban
managers and politicians allow the crisis to spiral out of control?
There was plenty of early warning. Why were decision makers unable to
solve the problem proactively? A possible answer, of course, is SD was
not yet one of their problem solving tools. This root cause was not in
the model, but yet the concept model produced excellent symptom behavior.

> Some authors do not use dynamic hypotheses while respecting strictly
> the SD
> paradigm. For instance R.G. Coyle never mentions dynamic hypotheses and
> builds his models only from the principle of causality, not
> necessarily root
> one because of the 'heap of tortoises'. He does not use either reference
> modes to build models but only as a way of comparing reality to the model
> behaviour once it is built.
Let's assume that the definition of ""dynamics hypothesis"" is a theory of
what is causing a problem's symptoms, or more broadly, a theory of what
is causing a system's behavior of interest. Then even though Coyle
doesn't use the term, his models are each a dynamic hypothesis, because
they are based on causes of system behavior.

The terminology doesn't matter to me, except for the fact that to
converse efficiently, we need standard terms with standard definitions.

Jean-Jacques - Thanks for pointing these things out. For me these were
very focusing observations.
Posted by Jack Harich <register@thwink.org>
posting date Wed, 19 Mar 2008 18:13:56 -0400
_______________________________________________
Jean-Jacques Laublé <jean-jac
Senior Member
Posts: 61
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jean-Jacques Laublé <jean-jac »

Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>

Hi Jack.

This is a very theoretical discussion, and I try to apply SD to very
concrete problems.
I think that to illustrate your ideas and to prove that they can have a
practical application, it would be a good idea to show an example even very simple
with a model (we must not forget that SD has a lot to do about modelling)
that shows how to use your ideas and what is the added value of doing it.
Building such a model would give a basis for this discussion.
Otherwise if one cannot build a model illustrating the point, this
discussion has no interest for me. I precise: for me, as it may have an interest for other
people.

I agree that a model is always an hypothesis but what I was writing is about
how models are built. The classical method works with reference modes and
tries to imagine hypothesis that explains these behaviour. Coyle does not
base the way of building models on these reference modes, arguing that it
does not make sense to use reference modes at the same time as basis for the
construction of models and as basis for their validation unless it is only
to verify that the first work has been correctly done and that there is no
error in the process. But it cannot be used twice once for the conceptual
building and twice to verify the same concept.

I prefer to work without explicit reference modes during the building
process, even if I can have them in the back of my mind, because reference
modes are often not available in practical work.
You may have completely new situations where reference modes can be
misleading and it is better to think about diverse scenarios. You may miss
reference modes because of unavailability of data or too high costs to get
them.

If I have reference modes and it is of course a great help, I will try to
use them as a validation tool.

About closed models, Coyle studies both closed and open models.
My problems are mostly and unfortunately open, although I recognize that
working with closed models is better for reasons of simplicity. But I must
adapt my models to both endogenous influence and exogenous ones.

I recognise too that this the way I think presently and that it may change
in the future.
I am not sure that Jay Forrester was really thinking about using dynamic
hypothesis when writing about the necessity of a model to reproduce reality.
Jay knows certainly better than us what he meant.
About your root causes it looks very much like boundary extensions.
I replace the heap of tortoises by a box in a box in a box in a box etc.

You write :
<This is because the field still clings to phrases
<like ""dynamic hypothesis"" of what is causing the problem instead of root
<cause and feels that if a model can reproduce the symptoms, then it must
<contain the root cause.

I do not think that anybody pretends that reproducing the symptoms does
guarantee anything.

The notion of dynamic hypothesis is not part of the SD paradigm but belongs
to the most known and taught method of building models.

Some authors do not use dynamic hypotheses while respecting strictly the SD
paradigm. For instance R.G. Coyle never mentions dynamic hypotheses and
builds his models only from the principle of causality, not necessarily root
one because of the 'heap of tortoises'. He does not use either reference
modes to build models but only as a way of comparing reality to the model
behaviour once it is built.

And I personally prefer much more this way of building models than the
official one.

Regards.
Jean-Jacques Laublé Eurli Allocar
Strasbourg France.
Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>
posting date Thu, 20 Mar 2008 10:38:19 +0100
_______________________________________________
Jack Harich <register@thwink.
Member
Posts: 39
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jack Harich <register@thwink. »

Posted by Jack Harich <register@thwink.org>
posting date Thu, 20 Mar 2008 09:30:32 -0400
_______________________________________________
Bill Braun <bbraun@hlthsys.co
Member
Posts: 43
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Bill Braun <bbraun@hlthsys.co »

Posted by Bill Braun <bbraun@hlthsys.com>

Jack Harich offers the Dueling Loop model as an example of root cause,
and high and low leverage points.

http://www.thwink.org/sustain/articles/ ... _Paper.htm

I am curious about the high leverage points; they are both exogenous
parameters (bottom of page on DuelingLoops_Paper.htm). I interpret this
to mean that the structure is just fine, we only need to pump different
content through it to achieve the results we want. Is that a fair and
accurate way of understanding the model?

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Fri, 21 Mar 2008 07:41:50 -0400
_______________________________________________
Jack Harich <register@thwink.
Member
Posts: 39
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jack Harich <register@thwink. »

Posted by Jack Harich <register@thwink.org>


SDMAIL Bill Braun wrote:
> Posted by Bill Braun <bbraun@hlthsys.com>
> I am curious about the high leverage points; they are both exogenous
> parameters (bottom of page on DuelingLoops_Paper.htm). I interpret
> this to mean that the structure is just fine, we only need to pump
> different content through it to achieve the results we want.

Bill,

Thanks. I don't understand what you mean by ""pump different content
through it."" What is ""it""? What is ""pump""? What is ""different content""?
Also, I don't know what you mean by ""the structure is just fine.""

I can say what the paper says. The highest HLP appears to be ""general
ability to detect political deception."" This is currently low, at an
estimated 20%. If it can be raised to a high level, 80% or above, the
race to the bottom strategies of falsehood and favoritism no longer
work. This causes the race to the bottom's dominance to collapse, as
Supporters Due to Degeneration flee for their lives to Supporters Due to
Rationality. This cause the race to the top to become the dominant loop.
This represents a possible solution to the change resistance problem the
paper discusses.

I'd be glad to try to explain the model and answer your questions. It's
just that I don't what to make any false assumptions. Sorry if the
writeup is less than complete or clear. It takes a lot to describe this
simple model, because it's a very abstract concept model and contains a
number of novel concepts.

Perhaps a good way to proceed would be for you (or others) to quote
passages from the Dueling Loops paper that you have questions about or
differ with. Or in some cases, there may be no such passage, such as an
implied premise or conclusion. The long version of the paper is the
better source for this.

Looking forward to a productive discussion,

Jack
Posted by Jack Harich <register@thwink.org>
posting date Sun, 23 Mar 2008 16:48:57 -0400
_______________________________________________
Bill Braun <bbraun@hlthsys.co
Member
Posts: 43
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Bill Braun <bbraun@hlthsys.co »

Posted by Bill Braun <bbraun@hlthsys.com>

Jack writes, ""Thanks. I don't understand what you mean by ""pump
different content through it."" What is ""it""? What is ""pump""? What is
""different content""? Also, I don't know what you mean by ""the structure
is just fine.""

I interpret the model to mean that the dynamic hypothesis is understood
to be a function of exogenous parameters, not current policies. By ""just
fine"" I meant that the policies explicit in the model are taken to be OK
as is, no changes are required or needed. By different content, I meant
that all that is required to achieve desired results is to change the
values of the exogenous parameters (which you identified as as the high
leverage points), which in the context of the model, means that people
only need to autonomously behave better than they are at present (their
behavior is an independent variable).

Alternately stated, I take Jack to be saying that (based on my
understanding of the model, and I may be wrong) the structure of the
problem focus is OK as is, and the root cause is defined as exogenous
parameters (having been identified as the high leverage points).

I am clarifying, not arguing, my point. I am still curious about my
interpretation of exogenous parameters as high leverage points. I can
think of any number of situations where it would be easy/convenient to
say, ""behave better"" and the problem would be solved.

Jack inquires into my post on root cause and exogenous variables. After
some thought, I think my question is generic.

To what degree ought we think of an exogenous parameter as a root cause
and/or as a high leverage variable, and correspondingly construct models
based on that?

If such thinking is sound, why build models in the first place? I think
I would be left with advising a client, ""Bill Braun, who is not under
your control or influence, is behaving badly. If you can get him to stop
(paradox intended), your problems are solved.""

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Sun, 23 Mar 2008 21:11:12 -0400
_______________________________________________
Jean-Jacques Laublé <jean-jac
Senior Member
Posts: 61
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jean-Jacques Laublé <jean-jac »

Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>

Hi Bill

You write
<To what degree ought we think of an exogenous parameter as a root cause
<and/or as a high leverage variable, and correspondingly construct models
<based on that?

An exogenous factor can be a root cause of the behaviour of a model, but can
hardly be a leverage point. A high leverage point is a factor that one can influence
in such a way that it will have a great influence on the behaviour of the
model with a minimum of effort. It is by definition endogenous and under control.

But it depends on the definition of a high leverage point.
High leverage point does mean; with a high influence or with a high
influence that is directly or indirectly under control which means that one has the possibility to
change the value of the factor?
For me a leverage point is a high influence factor that one can control.

Root causes can be exogenous or endogenous. If they are endogenous one can
try to influence them directly it they are constants under control.
Otherwise, whether the factor is endogenous out of control or not or
exogenous, the solution is to work with other leverage points to try to find
policies adapted to the influence of that factor.

In the case of the duelling loop that I have not the time to study (sorry
for that), if things are like you depict them, the model has no reasons to
exist and one has no or very little endogenous capability to thwart the
exogenous factors as there is no high endogenous leverage point.
The situation described is like a boat in a storm that has no rudder.

Your comments are then completely justified.
The model is completely contrary to the traditional SD school of thought
that considers mainly purely endogenously driven models. It is there a
nearly pure exogenously driven model.
Regards.
Jean-Jacques Laublé Eurli Allocar
Strasbourg France.
Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>
posting date Tue, 25 Mar 2008 09:38:06 +0100
_______________________________________________
""John Gunkler"" <jgunkler@sp
Member
Posts: 20
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by ""John Gunkler"" <jgunkler@sp »

Posted by ""John Gunkler"" <jgunkler@sprintmail.com>

Fred Nickols writes, about the example (from
http://www.systems-thinking.org/rca/rootca.htm) of the plant manager who
finds oil on the floor,

""I'll wager that aficionados of root cause analysis (RCA) would indicate
that the plant manager needed to go at least one or two steps further and
determine why he was pressing everyone to be so cost conscious.""

And I, being one such aficionado, might reply much in the light of Gene
Bellinger's suggestion at the bottom of the cited webpage: to wit, that
maybe we should not look for root causes but, instead, for ""actionable
causes"" that ""I can act on that will provide long term relief from the
symptoms, without causing more problems that I have to deal with tomorrow.""

In that light, if the manager can simply change his own behavior without
further analysis, and if that change in behavior resolves the problem
without creating more, then no further analysis is warranted. I notice,
too, that simply being put under pressure to reduce costs should not be
enough to prevent the plant manager from changing -- after all, if his bonus
depended upon cost savings, he has just discovered that certain actions led
to the opposite result and is fairly confident that a change in his behavior
would lead to a higher bonus.

If, on the other hand, the pressure on the manager was of a different sort
that prevented him from simply changing the way he pursued cost savings
(such as being held to very short-term accountabilities), then further
analysis might be useful and, indeed, another deeper cause and corrective
action might be preferable.

John Gunkler
Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
posting date Mon, 24 Mar 2008 12:46:46 -0400
_______________________________________________
Jack Harich <register@thwink.
Member
Posts: 39
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Jack Harich <register@thwink. »

Posted by Jack Harich <register@thwink.org>


SDMAIL Bill Braun wrote:
> Posted by Bill Braun <bbraun@hlthsys.com>
>
> Jack writes, ""Thanks. I don't understand what you mean by ""pump
> different content through it."" What is ""it""? What is ""pump""? What is
> ""different content""? Also, I don't know what you mean by ""the
> structure is just fine.""
>
> I interpret the model to mean that the dynamic hypothesis is
> understood to be a function of exogenous parameters, not current
> policies.
Hi Bill,

Have you read the Dueling Loops paper? I noticed that you are not using
any of the key phrases from the paper. Perhaps to save time you looked
only at the diagram titled The Leverage Points of the Dueling Loops at
the bottom of
http://www.thwink.org/sustain/articles/ ... _Paper.htm. I can
understand. We are all pressed for time.

On page 2 the paper says ""There are two feedback loops in the human
system that, in the large, affect citizen’s lives more than anything
else. They are the loops that politicians use to gain supporters."" Thus
the human system is the model boundary. Sterman defines exogenous
variables as arising ""from outside the boundary of the model."" (page 95)
Thus all the variables in the model are endogenous.

Perhaps you feel that constants or lookup tables are automatically
exogenous?

Regardless of the terms used, the key distinction that Sterman tries to
make is that one should not omit important variables from the feedback
influence of a model, if they are significantly affected by that feedback.

If you read the paper, you will see that the structure of the model is a
reflection of current policies, ones that are extremely common in
political systems of all kinds. Something is wrong in the current
policies. The model is an attempt to find out what.

> By ""just fine"" I meant that the policies explicit in the model are
> taken to be OK as is, no changes are required or needed.
I'm lost here. ""The policies explicit in the model"" are NOT okay. That's
the problem!

The purpose of serious problem solving models is ultimately to allow
decision makers to change their existing policies. Looking at popular
high quality models, like Forrester's urban decay model, the Club of
Rome's World3 model, and Jack Homer's and CDC's diabetes epidemic model,
we see that in all cases the problems are caused by bad policies. (A bad
policy includes no policy.) The purpose of these models is to help us
discover good policies. Not perfect policies, just ones good enough to
solve the problem.

> By different content, I meant that all that is required to achieve
> desired results is to change the values of the exogenous parameters
> (which you identified as as the high leverage points), which in the
> context of the model, means that people only need to autonomously
> behave better than they are at present (their behavior is an
> independent variable).
>
If you read the full length version of the paper, figure 18 on page 17
shows a subsystem that allows decision makers to ""push"" on the high
leverage point of ""general ability to detect political deception."" The
model in The Leverage Points of the Dueling Loops diagram is not the
final model. Figure 18 shows how this subsystem has been added, so that
the amount of Ability to Detect Deception is no longer a constant. It is
now a stock, with multiple inputs that affect its growth rate. The stock
is now in the path of model feedback loops.

Thus the model never implies that ""people only need to autonomously
behave better."" It shows that in the normal course of political system
behavior, people's ability to detect political deception goes up and
down. The key point, emphasized in the caption to figure 18, is that
""This simple subsystem imitates how society reacts when corruption rises
above an unwritten, culturally defined critical point. This reaction is
part of a cycle that never ends, because presently there is no formal,
enduring mechanism in governments to keep Ability to Detect Deception
permanently high.""

I apologize that the full length paper is so long, at 29 pages. But I've
found that a shorter presentation leaves out crucial details, and does
not allow key subjects to be covered in the depth they require.


> Alternately stated, I take Jack to be saying that (based on my
> understanding of the model, and I may be wrong) the structure of the
> problem focus is OK as is, and the root cause is defined as exogenous
> parameters (having been identified as the high leverage points).
>
> I am clarifying, not arguing, my point. I am still curious about my
> interpretation of exogenous parameters as high leverage points. I can
> think of any number of situations where it would be easy/convenient to
> say, ""behave better"" and the problem would be solved.
>
> Jack inquires into my post on root cause and exogenous variables.
> After some thought, I think my question is generic.
>
> To what degree ought we think of an exogenous parameter as a root
> cause and/or as a high leverage variable, and correspondingly
> construct models based on that?
>
> If such thinking is sound, why build models in the first place? I
> think I would be left with advising a client, ""Bill Braun, who is not
> under your control or influence, is behaving badly. If you can get him
> to stop (paradox intended), your problems are solved.""
>
Thanks for explaining. I feel your frustration. There are many business
and social problems that we'd like to see solved. In some cases we can
see a solution what would seem to work. But when we ""push"" on the system
to implement a preferred solution, little happens. We can usually see
this will be the case, so we don't bother to try.

Here's where I'm coming from. Every problem solver(s) has some small
amount of force they can exert on a system, in an attempt to change it
to preferred behavior. If their force is applied to a low leverage
point, solution failure is likely, because the force is too small. (The
force is the amount of effort to prepare and implement a change.) But if
they push on a point with high enough leverage, they will prevail. In
difficult problems, the HLPs are hard to find. This is made much easier
by finding the root causes first.

That's why we build simulation models: to soar past the constraints of
intuition.

Or, if one discovers a variable is exogenous and should be part of a
model's feedback loops, then it needs to become endogenous.

SDMAIL John Gunkler wrote:
> Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
>
> Fred Nickols writes, about the example (from
> http://www.systems-thinking.org/rca/rootca.htm) of the plant manager
> who finds oil on the floor,
>
> ""I'll wager that aficionados of root cause analysis (RCA) would
> indicate that the plant manager needed to go at least one or two steps
> further and determine why he was pressing everyone to be so cost
> conscious.""
>
> And I, being one such aficionado, might reply much in the light of
> Gene Bellinger's suggestion at the bottom of the cited webpage: to
> wit, that maybe we should not look for root causes but, instead, for
> ""actionable causes"" that ""I can act on that will provide long term
> relief from the symptoms, without causing more problems that I have to
> deal with tomorrow.""
>
Yes. That's why I define the characteristics of a root cause as:
(1) It is clearly a major cause of the symptoms.
(2) It has no productive deeper cause.
(3) It can be resolved. Sometimes it’s useful to include unchangeable
root causes in
your model for greater understanding. These have only the first two
characteristics.

The word ""productive"" was added as a result of this thread. The word
allows you to stop asking why at some appropriate point in root cause
analysis. Otherwise you may find yourself digging to China.

""Actionable cause"" is a nice phrase. This idea is incorporated in my ""It
can be resolved.""

Now then, what happens if it's too expensive to resolve directly? It
remains a root cause. You then have to resolve it indirectly. This can
be done by seeing that the root cause variable affects, and then
introducing forces that negate that influence.

> In that light, if the manager can simply change his own behavior
> without further analysis, and if that change in behavior resolves the
> problem without creating more, then no further analysis is warranted.
> I notice, too, that simply being put under pressure to reduce costs
> should not be enough to prevent the plant manager from changing --
> after all, if his bonus depended upon cost savings, he has just
> discovered that certain actions led to the opposite result and is
> fairly confident that a change in his behavior would lead to a higher
> bonus.
>
> If, on the other hand, the pressure on the manager was of a different
> sort that prevented him from simply changing the way he pursued cost
> savings (such as being held to very short-term accountabilities), then
> further analysis might be useful and, indeed, another deeper cause and
> corrective action might be preferable.
>
Yes. A fine educational example.

Thanks for this,

Jack
Posted by Jack Harich <register@thwink.org>
posting date Tue, 25 Mar 2008 18:53:17 -0400
_______________________________________________
Bill Braun <bbraun@hlthsys.co
Member
Posts: 43
Joined: Fri Mar 29, 2002 3:39 am

QUERY Definition of root cause

Post by Bill Braun <bbraun@hlthsys.co »

Posted by Bill Braun <bbraun@hlthsys.com>

Jack asks if I have read the Dueling Loops paper (not seeing me use any
key phrases) and notes that Sterman defines exogenous variables as
arising ""from outside the boundary of the model."" (page 95) Thus all the
variables in the model are endogenous. Finally Jack asks ""Perhaps you
feel that constants or lookup tables are automatically exogenous?""

I am taking the model at the bottom of Dueling_Loops.htm at face value.
I have not read the paper. My generic question is based on identifying
exogenous variables as high leverage points.

And perhaps my understanding of exogenous is dead wrong. I draw a
distinction between constants that represent behavior over which
decision makers have no control and parameters which are dimensions of
policy or decision making over which they have control.

I might use a constant to represent the economy if the time frame of the
model is short enough that changes in the economy are unlikely. The
economy is an exogenous variable. If I am looking at the time over which
I adjust inventory, that is a parameter and under my control. According
to my understanding (perhaps in peril here) all the variables in the
Dueling_Loops model are not endogenous.

I'm not sure how a lookup table could be exogenous. It either uses
another model variable or the implicit variable of time as its input
(the ""x"" in y=f(x)).

Returning to Jack's specific model, when he notes, ""I'm lost here. ""The
policies explicit in the model"" are NOT okay. That's the problem!"" then
I would expect to see one of those policies identified as a high
leverage point, the changing of which would change model behavior. (Or,
one model that represents current reality and another that shows
changed, added, or deleted policies.) Stated otherwise, I would expect
high leverage to include a feedback loop.

I sense the deeper question here is, am I on or off the mark in my
understanding of exogenous versus endogenous variables. I imagine this
to be tediously boring to people who do understand the difference, and
is my last comment on the topic.

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Wed, 26 Mar 2008 07:32:00 -0400
_______________________________________________
Locked