Stray dogs, street gangs and terrorists: manifestations ofa
latent capacity support system
(revised April 13, 2009)
Khalid Saeed
Professor of Economics and System Dynamics
Worcester polytechnic Institute
Worcester, MA 01609
ABSTRACT
This paper attempts to explain the inefficacy of public policies aimed at eradicating
unwanted activities. It is posited that such policies are based on a perceived
structure of the system that is not cognizant of the latent capacity support
mechanisms for the activity. The ineffectiveness of a policy can be explained when
the latent capacity support mechanisms are included in the model of the system.
Policies aimed at influencing the capacity support system would be far more
effective, compared with those directly attempting to contain the unwanted activity.
Introduction
Pests and other undesirable populations offer considerable resilience to
interventions attempting to limit or eradicate them. These interventions often
directly aim at reducing these populations through either by limiting their
procreation or accelerating their demise. For example, stray dogs roam the streets
of many major Asian cities in spite of the efforts to euthanize or castrate them.
Street gangs and the violence associated with them continue in may North American
cities in spite of considerable effort on part of law and order institutions to contain
them. Suicide attacks terrorizing public continue in central Asian countries like
Pakistan, Iraq and Afghanistan in spite of the concerted military offensives to
eradicate their perpetrators. This article suggests that a latent capacity support
structure that maintains the inflow into the undesirable population exists in all such
cases. Interventions not cognizant of this structure may only address symptoms and
Page | of 16
not alleviate the root cause. The capacity support process as a generic metaphor
may help to bring the latent root causes to fore.
Stray dogs as a manifestation of a latent capacity support system
My inspiration for defining a generic system for undesirable populations came from
observing the city government attempts to deal with the throngs of stray animals,
mostly cats and dogs, roaming the city of Bangkok. The local newspapers routinely
reported on the menace of the stray animals as well as the efforts of the city
government to deal with them.
General Chamlong Srimoung, was elected governor of Bangkok in the late 1985.1!
Chamlong came to office through a landslide victory and enjoyed wide popularity
among all cross sections of people due to his impeccable honesty and personal
charisma. A devout Budhist, vegetarian, and a declared celibate, Chamlong led a very
simple life. He dressed in a peasant’s clothes, lived in a bare one-room apartment
and slept on floor even when he was the governor of Bangkok. Chamlong sincerely
wanted to improve living conditions in Bangkok. Among other things, he also
wanted to alleviate the menace of stray dogs. However, being a kind man, he was not
in favor of euthanizing the animals as has been practiced in many cities of the world.
He wanted to find a humane solution to the problem. With help from the city’s
philanthropists, Chamlong set up a dog asylum. Stray gods were caught and brought
to this asylum, where they were castrated and treated for any ailments they had. A
small number of these castrated dogs were adopted, but most were released back
into the streets. This measure was expected to bring down the breeding rate of the
stray dogs and hence reduce their population, but experience shows this did not
happen. Elsewhere catching stray animals and euthanizing them is acommon
1 Chamlong Srimuang received the coveted Raymond Magsaysay Award in 1992. See
http://www.rmaf.org.ph/Awardees/Biography/BiographySrimuangCha.htm
Page 2 of 16
practice that often provides employment to dog-catchers without making a dent into
the stray animal populations.
Perceived structure of the stray dog population system
The perceived structure sustaining stray dogs clearly comprised a single
conservative system representing the stock of the undesirable population and its
growth and demise processes as shown in Figure 1.
fats] stray dog pop Ad
dog deaths
dog births
@%, ie
eo
3
life expectancy
fr births
Figure 1: Perceived structure of a pest population
The castration policy in this structure will limit the population by reducing its birth
rate; the euthanasia policy will attempt to achieve the same objective by increasing
the death rate. Both policies will be effective if there were no other flows or
influences affecting them were present. However, if there existed migration flows
into the population and the birth, death and migration rates were sustained by a
latent capacity support system, any intervention into limiting the birth and death
flows would be mitigated by the latent structure, which would render both policies
ineffective.
Latent capacity support system maintaining stray dog population
In fact, the gracious governor of Bangkok and his philanthropist friends did not
consider that the stray dog problem was not local to Bangkok, whereas the
castration program was. Stray dogs roamed all over the Kingdom of Thailand and
did not need a visa to walk into Bangkok.
Page 3 of 16
Bangkok, with its exotic sights, sounds and smells attracts millions of tourists every
year from all over the world. An important part of Bangkok is its restaurants and
food vendors that serve visitors with a variety of tastes and budgets. Unlike the
Western ethic, not finishing everything on your plate is considered polite and
tasteful so you are not considered a glutton. The left food is often fed to the stray
cats and dogs or just thrown away in a way that it remains accessible to the stray
animals. In a city of over 14 million, huge quantities of left food are generated.
Stray dogs are attracted to a place where they can feed well and Bangkok with the
same amount of food and a lower dog population created by castration would be a
favorite. Dogs may also spend more time breeding and less time looking for food if
there is plenty of food around. Needless to add that well fed dogs live longer than
the malnutritioned ones. So, any decreases in their population, whether from
euthanasia or castrations would soon be mitigated. These mechanisms create a
latent capacity support system for the stray animals that renders the policies to
address symptoms ineffective.
Figure 2 shows the complete model subsuming the latent capacity support structure
described above. Trash food availability determines population capacity in the long
run that drives immigration, births and deaths. The trash food availability in turn is
sustained by trashed food generation that creates a carrying capacity. Please note,
however, that carrying capacity is not given by the stock of trashed food, but by the
inflow into this stock. The long run carrying capacity would be zero no matter how
big the stock of trash food is if there is no inflow into this stock.
Figure 3 shows the performance of the castration and euthanasia policies, which are
effective only in the short run. Both improve food availability for the remaining dogs
in the city leading to increased in-migration from the neighboring districts, longer
life expectancy and more breeding time, all of which restore their local population.
Unless the latent capacity sustenance system created by trash food generation is
addressed, any policies directly attempting to limit their numbers will be defeated.
The only effective approach would be to limit the latent capacity by disposing the
Page 4 of 16
trash or reducing its generation, which is borne out by the additional simulations in
Figure 3.
fas) stray dog pop Ag
normal fr births dog births
dog deaths
Bangkok Ne)
stray dogs
2) life expectancy
normal life exp
ia) carrying capacity Ag
trash food
availability®
per dog
trash need
trash food —
generation
o—S— x9
trash food
consumption
by dogs
trash disposal
fraction
8 disposed
Figure 2: Complete system subsuming apparent and latent structure of a
pest population
Page 5 of 16
If we now consider that an improvement in the cleanliness of the city and reduction
in its stray pests might increase tourist inflow with its concomitant trash food
generation, even the trash disposal policy will not work unless there is a
fundamental change in the food throwing norm, which might be difficult to achieve.
It would seem that public policy problems created by latent mechanisms would
invariably resist interventions addressing symptoms. Interventions addressing
latent mechanisms sustaining the symptoms would be effective but then recognizing
the latent structure is admittedly not an easy task.
I Bangkok stray dogs: 1-2-3-4-5- Initial equilibrium
a=
—S—-
Euthanasia policy
astration policy 24 Trash disposal policy
5 54
s+
4 4
4
X 4
Reduced trash generation
4 8
1 35 T
0.00 5.00 10.00 15.00 20.00
Page 1 Years 3:48 AM Tue, Feb 10, 2009
NeeaF ? Untitled
Figure 3: Pest population response to various policy interventions
I have suggested in Saeed and Pavlov (2008) that developing an inventory of
metaphorical models addressed to specific episodes whose structure can be
extended to a whole class of problems might help to comprehend the latent systems
underlying them. This practice is consistent with the modeling process adopted in
operations research (Hillier and Libberman 1972) and also with Forrester’s position
Page 6 of 16
that there might exist a score or so generic systems that can be applied to a wide
range of problems we experience (Forrester 1980). I view the stray dogs model
suggested above to represent a metaphorical latent capacity support system that
may help in understanding the root causes in a wide range of situations involving
undesirable populations and activities. In the following section, I'll attempt to
extend this metaphor to address the problems of gang violence that continues to
plagues many North American cities and terrorism that continues to persist in South
and Central Asian countries in spite of the concerted efforts to eradicate them.
Extending the latent capacity support metaphor to street violence and
terrorism
Street violence and more recently terrorism are constant threats to humanity. Both
have also drawn much attention from law and order and defense institutions that
have often targeted elimination of leaders of these activities as a solution to the
problem. Presumably, the leaders are not only responsible for ordering gang killings
and terrorist acts, they also actively engage in recruitment of members in both
cases.? The leaders are also not easy to eliminate since they are often protected by
their own security infrastructure and considerable collateral damage may occur in
pursuing their elimination. The pursuit of the leader elimination policy has not
helped in curbing menace in both cases since a latent structure sustaining
recruitment of members into the system stays in tact. In this section, I’ll attempt to
outline this latent structure for street gangs and terrorists using the stray dogs
metaphor as a guideline.
Street violence
Human organizations are often driven by leadership, which emerges from within.
The menacing population in the human context can therefore be divided into
2 Street violence and terrorism have sometimes been viewed as analogous
phenomenon. See, for example: Berry, N., et al. (2004).
Page 7 of 16
member and leader categories that are a part of the same conservative system. The
perceived structure of this system is shown in Figure 4 for the case of street gangs. 3
3a perceived structure As
avg fractional kills
per leader per year
fraction unauthorized kills
fr leaders change in fr
eliminated leaders eliminated
leaders
eliminated
gang
vember readers
assuming leadership
killings
avg recruits per gang
member per year /desertion rate per 6
elimination i
perceived
avg recruits per collatoral threat
leader per year damage
~ (a) perceived threat
relative Hesernene pier per elimination
safety of leader attrition collatoral damage
gong per elimination
Figure 4: Perceived structure of street gang activity leading to a policy of
leader elimination
Gang members are recruited both by members and leaders, but the yield per leader
is much higher than the yield per member. Desertions from member ranks are
proportional to the gang population but are drive also by leader attritions, which
can occur after an average life of leadership or due to an active leader elimination
policy. Gang member eliminations occur through autonomous killings by gang
members as well as through leader authorized action. A small fraction of members
move to leadership ranks to balance leader attrition.
This system can be initialized in equilibrium. When a leader elimination policy is
introduced by stepping up fraction of leaders removed from the streets, the gang
3 Causal relationships underlying this model were suggested by my student Sarah
Boyer. See also Klein (1997).
Page 8 of 16
population will indeed assume a declining trend if this structure completely
represented the gang ecology. Experience shows however that simply eliminating
the leaders, who incidentally are also well-protected and not easy to reach, cannot
control gangs and their related violence.
There is clearly a latent structure hidden from the eye that neutralizes the
intervention to eliminate its leaders. I have added two computed variables -
collateral damage and perceived threat to the model of Figure 4 that might
precipitate change in gang structure as well as its environment. I have also included
an influence of the relative safety in gang on member enlistment. The two computed
variables seemingly invoke a chain of events in the latent structure that affect
recruitment through changing perceptions about the relative safety of affiliating
with a gang. Figure 5 shows the perceived structure of the latent system that
possibly sustains the gang in spite of the intervention to eliminate its leaders.
The relative safety of gang that encourages membership is a measure of the
differential of the gang security and the ambient safety. The former depends on the
gang security infrastructure, the later on the extent of collateral damage arising out
of gang operation. Gang security infrastructure is in turn created by investment for
maintenance as well as for enhancements invoked by increased perceived threat. It
decays due to entropy.
Oe) latent structure Aa
ambient safety relative: investment
© safety of per unit threat perceived
threat
collatoral
damage
gang gang security
infrastructure
normal
ambient
ae
safety gang investment
change in security in gang security leaders
bsorral seinen productivity of 10, ea infra
safety gang security lecay
infrastructure 8
Figure 5: Latent structure sustaining gang activity in the face of leader
elimination policy
Page 9 of 16
When gang leader elimination rate is stepped up, the ensuing threat forces the gang
to invest more in its security infrastructure, which enhances the relative safety in
gang. The collateral damage tied to leader elimination on the other hand reduces
ambient safety, which further enhances the relative safety of the gang. Enhanced
relative safety in turn facilitates recruitment. The intervention is therefore rendered
ineffective. The simulated behavior of the system subsuming both the perceived
structure of Figure 4 and the latent structure of Figure 5 is shown in Figure 6. The
model is started in equilibrium, which is disturbed by stepping up the fraction of
leaders eliminated.
s 1: gang members 2: killings 3: gang leaders 4: gang se...rastructure
I: 6004
2 110
3 35 ee
4 20
killings ——. |
Gang —
1: 350 .
2% 304 members
3 20
4 10 .
Gang security
infrastructure
q 4
1 100 \; Lea Gang leaders
2 10
3 5 _—
4 Op! T T T
0.00 15.00 30.00 45.00 60.00
Page 1 Months 4:30 AM Fri, Feb 13, 2009
a BaF ? Untitled
Figure 6: Model response to intervention to remove gang leaders from the
streets
Page 10 of 16
The intervention reduces the number of leaders in the short run, albeit it also
results in enhancement of the gang security infrastructure and reduction of ambient
safety that increase relative safety of gang that boosts its membership. As a result,
membership rises and the gang can easily replace its leaders. The gang size in fact
increases if nothing is done to build the ambient safety of the community.
On the other hand, when measures are taken to enhance the level of ambient safety
by community development and support mechanisms, the membership recruitment
process is stifled and the gang activity atrophies over time as shown in the
simulation of Figure 7. These simulations are borne out by experience in a number
of communities.
5.) 1: gang members 2: killings 3: gang leaders 4: gang se...rastructure
10041
20 ‘
10 Gang security
Infrastructure
wr
=
1: 50
2: 10
3 5 5 |
4 i] an. Gang leaders
killings
1 4
Gan
ie 0 8 2
2: 0} members “SY
4 0 1
0.00 15.00 30.00 45.00 60.00
Page 1 Months 4:44 AM Fri, Feb 13, 2009
a BaF ? Untitled
Figure 7: Model response to enhancement of ambient safety through
community development
Page 11 of 16
Terrorism
Terrorism by stateless actors has appeared in the last decade as a mega-menace that
threatens to evolve into a global conflict in spite of the concerted military efforts to
eradicate it. Learned opinions on the nature this menace have ranged from a clash of
major civilizations (Huntington 1996) to culture independent dissidence arising out
of sustained mistreatment and deprivation rampant in certain cross-sections of
communities (Sen 2006). Notwithstanding the value of these learned views, I wish
to surmise that the latent structure sustaining this global menace may not be much
different from that sustaining stray dogs in Bangkok or street gangs in North
America.
Figure 8 shows the perceived structure of the system that invokes concerted
military efforts. As in case of street gangs, the terrorists need leaders to function.
Recruitment is done both by the terrorists and their leaders, but the pool they
recruit from is more specific than that for street gangs. This pool consists of
dissidents often with personal agenda.
ET) perceived system Aa
avg fractional
suicides per
leader per year
fraction
autonomous
suicides
leaders
L
suicides
leader removal
ee)
ca
recruits assuming
leadership
fraction of
leaders removed}
avg fraction of dissidents
recruited per
terrorist per year desertion
rate
avg fraction of dissidents
recruited per z 8 change in fraction
leader per year dissidents fractional desertions of leaders removed
per leader removal
Figure 8: Perceived structure of terrorist activity
Page 12 of 16
Like street gang members, some terrorists may desert the group when a leader is
eliminated. They may also engage in self and member killings through organizing
suicidal acts that terrorize the public and kill innocent citizens. These acts are
largely undertaken at the instance of the leaders, but some autonomous suicides
also occur. The system of Figure 8 would collapse if leaders are eliminated - an
expectation that has driven the massive military actions against terrorism, since this
would stop suicidal actions, stifle recruitment and increase desertions.
Unfortunately, such military actions have hardly made a dent in the phenomenon.
Instead terrorist activity seems to have expanded as the military actions to
eliminate the leaders increase. There is clearly a latent structure that sustains this
activity.
Since terrorists are recruited from a population of dissidents, it would be important
to know how the latter are created. Figure 9 shows a simple view of the latent
structure feeding the stock of dissidents.
collatoral deaths
per removed leader
dead
8 latent system Ad
survivors per as
collatoral dead ractional
dissident attrition
leader removal ee J De)
urvivors of dissident
collatoral attrition yeeiis
collateral
deaths dissidents
recruited as
terorists
Figure 9:
Latent structure supporting recruitment of terrorists
The, often vengeful, survivors of the bystanders killed when the authorities pursue
the terrorist leaders feed the dissidents stock. Since leaders are not easy to reach,
this collateral damage can be frightfully high. The dissidents are either recruited as
Page 13 of 16
terrorists or eventually leave the dissident group if their vengeance is diffused
which can happen if they are helped to reconstruct their lives or even as time
passes.‘ The policy to kill leaders can therefore regenerate dissident population as is
borne out by the simulation of Figure 10 as well as by experience.
1
dissidents el
> aa
1
See
® >»: dissidents 2: leaders
1
4
1 150
2 6J
1 120
1 1
1 woe
aa suicides
1: 109 Sz a4 ee ee
2 = a
3 5 3 e leaders
4 +! T T T 1
0.00 15.00 30.00 45.00 60.00
Page 1 Months 9:06 AM Sat, Feb 21, 2009
ae 2? Untitled
Figure 10: Terrorist system response to leader elimination intervention
On the other hand, any interventions that help the attrition from the dissident group
may create a downward spiral in the terrorist activity as shown in Figure 11. Such
attritions can be facilitated by helping the populations affected by the collateral
damage to reconstruct their lives or by engaging these populations in constructive
economic activity as has been suggested by many observers.
4 See also Skarin (2002)
Page 14 of 16
& 1: dissidents 2: leaders 3: terrorists 4: suicides
13 10547—~2
5: 108
4: 6 RR,
1 3, leaders
4 Ny 2
1: 80
2 8 . ,
3} 80) terrorists “2.44
4: 4 2
i suicides
4 3
1
1: 55| dissidents
2 6
z 55
4: 2 T T T 1
0.00 15.00 30.00 45.00 60.00
Page 1 Months 9:30 AM. Sat, Feb 21, 2009
aaf ? Untitled
Figure 11: System response to interventions for diffusing dissidence
Conclusion
A pervasive latent structure providing capacity support to undesirable social
phenomenon often renders interventions directly attacking such phenomena
ineffective. Stray dogs, street gangsters and terrorists cannot be eliminated by
directly killing the perpetrators. Their capacity support mechanisms that are often
hidden from the eye must be recognized for designing effective intervention to
eradicate them. The real challenge is to bring to fore the latent structure enabling
continuation and sometime further growth of the undesirable activities. In all cases,
an investigation into motivations of the populations for engaging in the undesirable
activity will often lead to the root cause if the metaphor of a latent capacity support
system is kept in view. Other examples of activities that have resisted interventions
based on truncated perceived structure include crime, alcoholism, drug addiction,
overuse and depletion of commons, and even the uncontrolled growth of human
population. The metaphor of latent capacity support implicit in the stray dogs story
Page 15 of 16
can help understand the root causes of the resistance to interventions for reversing
growth trends in undesirable activities.
References
Berry, N., et al. 2004. Computational Social Dynamic Modeling of Group Recruitment.
Sandia National Laboratories report SAND2003-8754.
Forrester J W. 1980. System Dynamics--Future Opportunities. In: A. A. Legasto, Jr. et.
al. System Dynamics. North-Holland: New York. 14: 7-21.
Hillier, F S and Lieberman, G J. 1972. Introduction to Operations Research. San
Francisco, CA: Holden-Day, Inc.
Huntington, S. 1996. The clash of civilizations and the remaking of world order. Chicago,
Il: University of Chicago Press.
Klein, M. W. 1997. American street gang, its nature, prevalence and control. Oxford
University Press.
Saeed, K. and O. Pavlov. 2008. Dynastic cycle: A generic structure describing resource
allocation in political economies, markets and firms. Journal of Operations
Research Society. 59(10): 1289-1298.
Sen, A K. 2006. Identity and violence: the illusion of destiny. New York, NY: Norton &
Company.
Skarin, B. 2002. Understanding the driving factors of Terorism. Senior project.
Worcester, MA: WPI, SSPS Dept.
Page 16 of 16
