Dynamics of Reconstruction Projects
James R. Enos, Major, U.S. Army
Department of Systems Engineering
Mahan Hall, West Point, NY 10996
(845) 938-3114
james.enos@ usma.edu
Abstract
Ongoing counterinsurgency operations in Iraq present a complex, dynamic environment
in which traditional analytical methods struggle to explain the behavior over time. System
Dynamics is extremely well suited to analyze this environment as the methodology focuses on
understanding the structure of the system and the behavior it creates. This paper proposes a
system dynamics model of reconstruction projects for essential services to examine one aspect of
this operating environment. One of the many challenges that exist in this environment is
determining the proper balance between the use of the Commander’s Emergency Response
Program (CERP) for small scale and major reconstruction projects. This paper attempts to
mitigate this challenge by analyzing the structure of the system, modeling the behavior of the
system over time, and proposing policy recommendations to improve the system behavior.
Although the model is not calibrated to historical data, it produces behavior consistent with
behavior described in Army doctrine. The causal relationships provide valuable insights into the
dynamic behavior of reconstruction efforts and their impact on essential services. With further
calibration of the model, leaders can develop and evaluate policy alternatives for capacity
development to mitigate the impact of the insurgency.
Key Words: System Dynamics, Reconstruction, Counterinsurgency, CERP Funds
Introduction
Current operations in Iraq and Afghanistan have identified shortcoming in traditional
analytical methods for dealing with the complex, dynamic combat environment our military
faces today. A challenge that exists for commanders is determining the proper balance between
the use of the Commander’s Emergency Response Program (CERP) for small scale
reconstruction projects and major reconstruction projects. This paper proposes an initial system
dynamics model of reconstruction projects based on a literature review, interviews with subject
matter experts, and current Army doctrine. The model is then used to evaluate different levels of
reconstruction funding and the proportion of funds spent on large and small projects. With these
insights, leaders can develop policy alternatives for capacity development within the host-nation
to mitigate the impact of the insurgency.
Background
Since March of 2003, the Unites States has appropriated $61.64 billion for reconstruction
efforts in Iraq (Special Inspector General for Iraq Reconstruction 2011). Of that amount, only a
small fraction, $3.85 billion, is allocated to the Commander’s Emergency Response Program
funds. CERP funds began as a tool for commanders to stabilize their area of operations through
Page | 1
the initiation of emergency relief and small scale reconstruction projects. Initially, funds were
allocated from seized Ba’athist Party funds; however, judge advocates later ruled that
Department of Defense (DoD) funds could be used for these projects (Martins 2005). Initially,
commanders were able to spend up to $100,000 on reconstruction assistance defined by a
Combined Joint Task Force order as:
“the building, repair, reconstitution, and reestablishment of the social and material
infrastructure in Iraq. This includes but is not limited to: water and sanitation infrastructure,
food production and distribution, healthcare, education, telecommunications, projects in
furtherance of economic, financial, management improvements, transportation, and
initiatives which further restore the rule of law and effective governance, irrigation systems
installation or restoration, day laborers to perform civic cleaning, purchase or repair of
civic support vehicles, and repairs to civic or cultural facilities ” (Commander, CJTF-7 2003)
As the war progressed, the initial limit of $100,000 increased to incorporate larger scale projects
focused on developing essential services within Iraq.
Initially, the reconstruction efforts funded with CERP money became an outstanding
success as more than 11,000 projects were completed between June and October 2003 (Martins
2005). These projects spanned the range of providing school supplies to children to cleaning
major water supply systems. However, CERP Funds still comprise a very small portion of the
overall reconstruction effort in Iraq to improve the essential services the country provides for its
citizens. The System Dynamics model in this paper examines potential alternative policies
which increase the percentage of funds allocated to the CERP. It attempts to determine if there is
a tipping point at which allocating a different percentage of CERP funds to reconstruction efforts
will improve the essential service level in the host nation.
Literature Review
The Army’s new Field Manual 3-24, Counterinsurgency, provides an overview of
modem insurgencies and counterinsurgency techniques to determine the underlying dynamics of
reconstruction efforts. The manual presents the
argument that the legitimacy of a nation is the %e
underlying goal in any insurgency or ‘
counterinsurgency. Nation states attempt to gain
legitimacy by improving their economic situation,
essential services, and security within their borders.
Additionally, they attempt to minimize the level of
corruption to a level that is culturally acceptable to ==.
the people of the nation. In order to have a lasting, =
stable government it must be viewed as legitimate o 2 4 6 8 10
to a majority of the populace (Department of the Time
Army 2006). irate aadlal
FM 3-24 also provides a theory as to how an ke ‘i Haserill 13 “ical ¢ Wepuirtinet
increase in the quality of essential services provided of the Army 2006) SENIE ADAGE ep ern
by the host nation will impact the counterinsurgency ——
and insurgent effectiveness. Figure presents an adaptation of the behavior over time FM 3-24
Qualiyt of Services
Page | 2
describes. As shown, with an increase in essential services, the insurgent effectiveness decrease
at an increasing rate, similar to goal seeking behavior as the effectiveness of the
counterinsurgency increases. This would suggest that essential services have a negative link to
insurgent support and a positive relationship to the government’s legitimacy. Additionally,
Kilcullen describes an insurgency as a complex system that needs energy, in the form of acts of
violence and grievances against the government, to sustain itself through several feedback
structures (2004). This theory supports FM 3-24 assertion that an increase in essential services,
which reduces the grievances against the government, will act to decrease the effectiveness of
the insurgency. FM 3-24 provides additional insights into the behavior of reconstruction efforts
and the components of the system dynamics model.
System Dynamics is a methodology to understand the dynamic behavior of complex
systems through modeling and simulations. System Dynamics explains the behavior of systems
over time as a direct result of the system structure. It also aims to adjust individuals’ mental
models of the system to implement policies to improve the system. Forrester described the
potential for system dynamics as an approach that should help in the important high-level
management problems (1961). He noted that solutions to small problems will only yield small
results and that people get mediocre results by setting improvement goals too low. He suggests
that the change must be at the enterprise level to achieve major improvement and that the goal
should be to determine policies that lead to greater success (Forrester 1961).
Forrester describes a system as “a grouping of parts that operate together for a common
purpose” (1968). He further classifies two types of systems: open systems, in which exogenous,
or external, variables affect the system, or closed systems, in which all variables are endogenous,
or internal to the system (Forrester 1961). The distinction between open and closed systems
relies heavily on where the system boundary is drawn; however, a model of a system will
provide a better understanding of the dynamics the closer it is to a closed system. Dynamics are
the behavior of a system over time, which are generally complex and non-linear in nature
(Forrester 1961). This complexity comes from feedback within the system, time delays between
decisions and effects, and the learning process of the system (Sterman 2000).
Causal loops diagrams are a key element of the system dynamics approach which are
signed diagrams that represent the reinforcing or balancing feedback within a system. Casual
loops are different from discrete, event-oriented perspective of individual causes and effects in
that they acknowledge that in a closed system any cause is an effect and any effect is a cause
(Richardson 1991). In System Dynamics, the feedback loop diagrams indicate that one variable
influences another through physical or information flows. One is able to describe the behavior of
the system by talking through the loop to tell the story of the interactions within the system
(Meadows, Randers and Meadows 2004).
System Structure
The structure of the reconstruction effort system is extremely complex with several
individual feedback structures contributing to the dynamic behavior of the system. To
understand the structure of this system, this paper presents three of the major feedback structures
to include the desire to begin reconstruction projects, the security situation, and the maintenance
of essential services. The first feedback structure presents how degradation in essential services
will create the desire to begin reconstruction projects and how a population will desire a higher
level of service based on their current situation. The second feedback structure describes how
Page |3
the security situation in the country impacts the project completion rate, corruption, and projects.
The final feedback structure presented examines the impact of the difference in maintenance
complexity between large and small projects and how this complexity leads to a decrease in
essential services.
The first feedback structure within the reconstruction project system includes a balancing
and a reinforcing feedback loop. First, the “Need for Projects” loop explains how projects will
be started as a gap in essential
Ra services is felt by the nation. If
Projects Essential the level of Essential Services
Sombleted Bemiees decreases, then the Essential
An) mesic’ Ek r Services Gap will also increase.
B dawvor, EssentialSenices This will create a desire to
Need for Projects Want More improve the current situation by
increasing the Projects Started.
As more projects begin, the
number of Projects Completed
will also increase and then have
the desired effect of increasing
the level of the Essential Services. However, there is an interesting dynamic in that as people
experience a certain level of Essential Services they will increase their Desired Level of Essential
Services. This is shown in the “Have More, Want More” reinforcing loop. This loop creates the
requirement for additional reconstruction projects as the population begins to expect an increases
level of service.
The second feedback structure focuses on the security situation in the country and the
feedback with the reconstruction projects. This model does not attempt to model other
counterinsurgency methods, such as direct action against insurgents, instead focusing on the
impact reconstruction projects have on the
Essential
Projects Started Senlces Gap
+
Figure 2: Desire to Start Projects
overall system. The “Attacks on Projects” Cormuption GC)
feedback loop explains how insurgents can Comuption 4
directly impact reconstruction projects by Essential Legitmacyof
attacking the projects themselves. If there ages sovemne
is an increase in Attacks on Projects, the
number of Projects will obviously £ spas be) ‘
decrease. This decrease also causes a 2 Security
lower level of Essential Services, which project etonenees
decreases the Legitimacy of Government Somecton te)
and the overall Security in the country. As woieercrear Attacks on nace”
the security level decreases the overall Projects 4 ___.- Violence
number of Acts of Violence increases as =
does the number of Attacks on Projects Availity of
(Choucri, et al. 2006). So, if this Workforce
reinforcing loop begins to act in a negative Figure 3: Security Situation
manner it can have a profound impact on the essential services in the country. Additionally, this
can impact the Availability of Workforce as the workforce may become intimidated by insurgent
activity, which will decrease the Project Completion rate. Additionally, as the Legitimacy of
Government is decreased the “Corruption” feedback loop may begin to impact the behavior of
Page | 4
the system. Corruption within the country could decrease the Project Completion rate, which
would then decrease the number of Projects and eventually decrease the Legitimacy of
Government even further.
The final feedback structure explains the dynamics of the impact the different sized
projects have on the maintenance complexity and the eventual degradation of essential services
due to a lack of maintenance. The first feedback loop is a reinforcing feedback loop for “Large
Project Maintenance.”
This loop describes how an
increase in the Essential + pasion eet 2
Services Gap creates an + a 4, Sinal Prélects:
; 7 Desired Level of Essential
increased Desire to Start EssentialSerices Services Gap Large Projects An)
Projects, which leads to :
. Small Project
more Large Projects. Maintenance
However, with more AR) +
G z, Essential Large Project Maintenance
projects come an increase Services Mamnensnes Complexity “+
in the Maintenance -
Complexity and an increase
in the Maintenance Gap
Degregation of A +
between the required level EssenialSenicgs Malan
of maintenance and the +
. “Ts ;
Maintenance Ability of the ENnacyst
country (McDonald 2011). Government Maintenance Ability
As the Maintenance Gap NL wainerance
increases, there is a . = Ability
D egradation of Essential Figure 4: Maintenance Feedback Structure
Services which decreases the level of Essential Services and widens the Essential Services Gap.
Thus, this loop reinforces the behavior to start additional projects. This is also visible in the
“Small Project Maintenance” loop; however, not to the same extent as in large projects because
the complexity of individual projects is much smaller. An additional reinforcing loop exists in
the “Maintenance Ability” of the country. As the level of Essential Services increases, the
Legitimacy of Government also increases. With a more legitimate government, the nation is able
to increase its Maintenance Ability as is has resources to maintain the complex, large projects
and decreases the Maintenance Gap.
System Dynamics Model
The system dynamics model of the reconstruction projects is composed of three different
views; the essential services, security situation, and maintenance views. The feedback structure
described above provides the basis for developing the model and the relationships between the
different variables. The main components of the essential services view are the number of large
and small projects and the level of essential services in the nation. In the security situation view,
the major components include the level of government legitimacy, the security situation and acts
of violence. Finally, the maintenance view includes the required level of maintenance
complexity and the ability of the nation to conduct this maintenance.
The system boundary only included the reconstruction effort to improve the essential
services within the nation. This includes both large scale projects and CERP projects, which the
Page |5
model represents as small projects, aimed at developing the country’s essential services. Within
the security view, the model does not include any kinetic counterinsurgency operations aimed at
reducing the number of acts of violence. A more robust model could include kinetic operations
as a potential leverage point to further improve system performance; however, it was not
necessary to demonstrate the problematic behavior. Also, the model does not include an
insurgent growth element to demonstrate how insurgents are recruited and trained, but just
associates acts of violence with the security situation in the nation.
Time to Compete ee es
SmallProjects Desired Smal S
Project Stare mee
SP Approval—
Average Smat—_ ‘ime
Proectcost
Sralfiopes
Completed
Smal Projects
Possible small Sure
SP wes
. % * tee : os
Budget s NormalTime for BS
t* Nomat% of Small Sit] seat penne \ oe — ES Degregaton \
\ Propets not completed —_—- pleted .
{saya 7 Y | oped lesen em \
ato a } | Settee r meee \ |
\ « a ey | \ PN \ |
$e Albcated to | \ { Ss . 4
\ Gero / \ \ ~ a
u i \ ; hy
ee \ + % faite }/
|
vest seniet
| atest /
Sprague
LZ . v
LPwES \ reas Essential
se y | /
Total Large ie Ese! ES Desire
Ut i ee —
scones
LP approval x Time Complete
Time ~~) uatge Projects Large Projects
| 4 ne
\ (
Sai at \ ne |
ee : -
onan L j a
} =
oesiliine 4 ———
ek
Figure 5: Essential Services Model View
Figure 5 presents the essential services model view of the overall reconstruction effort
and includes the leverage points of the total budget and percent of funds allocated towards small
projects. The structure of this portion of the model is a co-flow, which depicts an increase in the
essential service level with the completion of both large and small projects. Also, the model
accounts for any projects, both large and small, that are no completed due to insurgent attacks on
the projects, intimidation of the work force, and corruption. These are all variables that can
impact the overall project completion rate for any reconstruction projects (Special Inspector
General for Iraq Reconstruction 2011). Additionally, the model accounts for the degradation of
essential services over time due to normal degradation of projects, which can be impacted by a
lack of maintenance ability and insurgent attacks against the infrastructure. The model also
accounts for the population’s increase in their desired level of essential services, as the higher
this level, the more services they will come to expect from the government. So, this reinforces
the number of projects that will be started to improve the nation’s essential services.
Page |6
Figure 6 presents the security situation view of the model, which includes the legitimacy
of the government as the major stock in this portion of the model. The legitimacy of the
government can change as the perception of the security situation and the essential services
increases (Department of the Army 2006). Both of these variables will impact the normal
change in legitimacy to determine the actual change in legitimacy. Like the essential services
view, the model accounts for the growth in the desired level of legitimacy as the population will
begin to desire a higher level of legitimacy as time progresses. The level of legitimacy of the
government impacts the support for the government, which then drives the security situation in
the country. Based on the level of security in the country, the model generates a number of acts
of violence, which can be targeted against infrastructure, reconstruction projects, or the work
force (Latawski 2006). Acts of violence against counterinsurgency forces were not included in
this model as they did not directly impact the problematic behavior observed.
Figure 6: Security Situation Model View
Figure 7 presents the final view of the model, which includes the maintenance gap
generated by a required level of maintenance complexity and the ability of the country to
maintain essential services. The model asserts that there is a level of maintenance ability, based
on the legitimacy of the government and the initial level of maintenance capacity of the nation.
This is the country’s ability to maintain infrastructure based on their education level, financial
situation, and work force experience. The maintenance gap exists as projects add complexity to
the infrastructure. It looks at the total percentage of large scale projects, which have a higher
level of maintenance complexity, and adds this value to the maintenance complexity of small
scale CEPR projects, which are much easier to maintain and have a lower maintenance
complexity (Special Inspector General for Iraq Reconstruction 2007). The maintenance gap then
increases the degradation of essential services because of a lack in maintenance.
Page |7
Laverne ——
umcate Pestle to
,
SE widen \
yf neni ~ \
. So] saline
[Completed Maintenance —— ‘duet lack of
Maintenance
oy :
Ls \reeena ! wt
‘Smal Projects | impacto
| Maintenance Gap —
SmallProject’ | zhi
Se Maintenance | Es:
Maletenance —— /
Level ofNation *
Ability to Mairi
Essemal Senices
Figure 7: Proj ect Maintenance M odel View
Results of Proof of Concept Model
Without historical data to calibrate the model to, the initial simulation attempted to
generate the behavior FM 3-24 presented. FM 3-24 proposes that as the essential services
increase, the effectiveness of the counterinsurgency also increases and the effectiveness of the
insurgency decreases (Department of the Army 2006). Figure 8 and Figure 9 present the
essential services and the legitimacy of government output from the initial simulation. As shown
in these figures, both are increasing after an initial decrease as the insurgency begins. This
appears to be consistent with the behavior FM 3-24 presents for this aspect of an insurgency.
Essential Services Legitimacy of Govemment
1,500 100
g 125 B
Essential Servions
Legitimacy
Pot anes Bi
375 le 5
let
0 hr TCL 0
o i 2% 3% 4 50 60 70 80 9 100 o 0 2% 39 4 50 6 7 0 90 100
‘Time (Month) ‘Time (Month)
Essential Services : Base Run. 343-3 4 pp Legitimacy of Goverument : Base Run
Figure 8: Essential Services - Base Run Figure 9: Legitimacy of Government - Base Run
The model does not have a specific variable for the effectiveness of the insurgency;
however, the two variables Acts of Violence and Corruption serve as reasonable approximations
of the insurgency’s effectiveness. These are also similar to the variables Kilcullen presents in his
paper as the requirements for an insurgency’s success (2004). Figure 10 and Figure 11 present
the acts of violence and corruption outputs from the base run of the simulation. Although these
values initially increase, after about 10 months, they begin to decreases over time. Again, the
model appears to generate a similar behavior to that of the theory in FM 3-24.
Page | 8
Acts of Violence Corruption
2» 4
15 re
he
mann Le
H j
5 Oa
et! | aL +
) tes) 0
o 0 2% 3 40 60 60 70 80 %0 100 o i 2 30 40 60 60 70 80 0 100
Time (Month) Time (Month)
‘Acts of Violence : Base Run + ‘Comuption : Base Run
Figure 10: Acts of Violence - Base Run Figure 11: Corruption - Base Run
Although a thorough calibration of the model was not completed, the model generates
behavior similar to that FM 3-24 proposes for the interaction of essential services, counter
insurgency effectiveness, and insurgency effectiveness. For the purposes of this paper, the
general trends in behavior provide insights into the performance of the system and potential
policies for the application of reconstruction funds. Additionally, the model generated the
predicted behavior during the extreme conditions tests. If the total budget decreases to about
$10,000,000 the insurgency begins to dominate the system and the level of essential services
decrease to a point in which it cannot recover. Also, the simulation does not improve past a
point when additional funds are allocated to the total budget for reconstruction efforts.
Potential Policy Recommendations
Although a more detailed, calibrated model would be required to develop policies, this
model demonstrates the predicted behavior for policy changes. The two major leverage points in
the model are the percent of the projects that are small projects and the total budget variables.
Changes to these two variables provide the inputs for the analysis of different policy alternatives.
The first altemative analyzed the impact of increasing the percentage of small projects from 5
percent to 20 percent. The second alternative examined the impact of increasing the percentage
of small projects to 35 percent and decreasing the budget to $300,000,000 per month. This
reduces the overall budget for reconstruction projects by over 50 percent.
Essential Services Legitimacy of Govemment
o 0 2 3 4 30 6 7 80 9 100 o wm 2 3 40 30 60 70 8 90 100
‘Time (Month) ‘Time (Month)
pe
sitettie —3 3 3 3 Fy wth 3
‘sential Services - Alternatives Figure 13: Legitimacy of Government - Alternatives
Figure 12 and Figure 13 present the essential service level and the legitimacy of
government level for the two alternative policies and the base run. As shown, both altematives
generate improved performance from the base case and create the same pattern of behavior over
time. Line 2 represents the increase in the percentage of small projects alone while the line 3
depicts the performance over time when the percentage of small projects increases and total
budget decreases. The interesting dynamic in this system is that both alternatives are capable of
generating the same pattern of behavior. So, the better alternative would be to decrease the total
budget by over 50 percent and to increase the percentage of small projects to 35 percent. This
increases both the essential services and legitimacy of government to a higher level than the base
run and will cost less to implement the reconstruction efforts.
Acts of Violence Corruption
CComsptiony Month
o 10 2 3 4 50 60 7 8 90 100
o 0 2 3 40 50 60 70 8 9 100 “Time (Month)
‘Time (Month)
reoBudget >
orruption - Alternatives
of Violence - Alternatives
re 14: Acts
Figure 14 and Figure 15 present the model output for the acts of violence and the
corruption, respectively, for the two alternative policies. As shown, both policies demonstrate
improved performance from the baseline. This output reinforces the policy recommendation do
both increase the percentage of small projects and decrease the total budget for the reconstruction
policies.
Conclusion
This paper presented a proof of concept model and initial simulation results of the impact
reconstruction projects could have in a Counter-Insurgency environment. Although the model
functions and provides output similar to the theoretical models that FM 3-24 provides, it is not
calibrated to historical examples of counterinsurgency. Future research and work could be
devoted to exploring how the model’s outputs compares to these historical examples given those
circumstances.
The major insight from this application of System Dynamics is that a better balance
between the number of small projects funded from the Commander’s Emergency Response
Program and the number of large reconstruction projects could potential reduce the amount of
funds required for reconstruction. Although the model only generated similar behavior to that
FM 3-24 presents as typical behavior in an insurgency, the insights appear to be valid. Future
work could include modifying the model to make more variables endogenous to the system.
Additionally, a more detailed calibration of the model could be conducted to determine the exact
funding levels and percent allocation to small projects to maximize the development of essential
services in a given nation.
Page | 10
References
Choucri, Nazli, et al. "Understanding & Modeling State Stability: Exploiting System Dynamics."
MIT Sloan Working Paper, January 2006.
Commander, CJTF-7. "FRAGO 89." Combined Joint Task Force 7, June 19, 2003.
Department of the Army. FM 3-24 Counterinsurgency. Washington D.C.: Department of the
Amny, 2006.
Forrester, Jay W. Industrial Dynamics. Waltham, MA: Pegasus Communications Inc, 1961.
—. Principles of Systems. Cambridge, MA: Wright-Allen Press, Inc., 1968.
Kilcullen, David. "Countering Global Insurgency." November 30, 2004.
Latawski, Paul. Tackling Counter-Insurgency and Post-Conflict Recconstruction: Recent
Experiences and Best Practices. Wilton Park Conference, 2006.
Martins, Mark S. "The Commander’s Emergency Response Program." J oint Force Quarterly
(National Defense University Press), no. 37 (2005): 46-52.
McDonald, Kenny, interview by James Enos. LTC (May 1, 2011).
Meadows, Donella, Jorgen Randers, and Dennis Meadows. Limits to Growth: The 30-year
Update. White RiverJunction, VT: Chelsea Green Publishing Company, 2004.
Richardson, George P. Feedback Thought in Social Sciences and Systems Theory. Pennsylvania ,
PA: University of Pennsylvania Press, 1991.
Special Inspector General for Iraq Reconstruction. "Quarterly and Simiauunal Report to the
Unites States Congress." Arlington, VA, 2011.
Special Inspector General for Iraq Reconstruction. "Quarterly and Simiauunal Report to the
Unites States Congress." Alexandria, VA, 2007.
Sterman, John. Business Dynamics: Systems Thinking and Modeling for a Complex World.
Boston, MA: Irwin McGraw-Hill, 2000.
Page | 11
About the Author
Major James Enos is currently an instructor in the Department of Systems Engineering at the
United States Military Academy, West Point, NY. Throughout his military service, he has held
numerous leadership positions as an infantry officer, including Rifle Company Commander,
Ranger Instructor, and Platoon Leader. He has spent almost 24 months deployed to Iraq as an
Infantry officer, where he managed numerous reconstruction projects and gained valuable
insights into problems addressed in this paper. He graduated from the US Military Academy at
West Point in 2000 with a Bachelor of Science degree in Engineering Management. He earned
his Master's of Science in Engineering and Management in 2009 from the Systems Design and
Management program at MIT. He teaches classes in modeling and simulation, systems
engineering, and system dynamics.
Page | 12
