Organized Crime and Economic Growth:
a System Dynamics Approach to a Socio-economic Issue
Vittorio Raimondi
University of Bergen
Information Sciences Department
Fr. Meiltzersgate 11b - 5007 Bergen - Norway
Mobile: ++47-414-63826
Email: vittorio.raimondi@ifi.uib.no / vittorio_r@ hotmail.com
Organized Crime and Economic Growth:
a System Dynamics Approach to a Socio-economic Issue
Abstract
This paper describes the potential of system dynamics models to support policy decisions and macroeconomics
strategies aimed at reducing the level of organized crime in a country and promoting further economic growth. A
causal loop diagram will elicit the main links found between organized crime, investment safety expectations and
economic growth. Additionally, it will suggest the need to focus the attention on what has been defined as the “risk
fraction”, that is unemployed people which constitute the recruitment base for crime organizations. This concept sets
the groundwork for the main dynamic hypothesis of a system dynamics model. Central in the structure of the model is
the attempt to identify the most relevant variables defining the “crime attractiveness”, to emphasize a dynamic
conceptualization of “risk fraction”, and to suggest alternative ways to quantify and evaluate the effectiveness of
various socio-economic policies.
KEYWORDS: organized crime, perceived investment safety, public policy, crime attractiveness,
system dynamics, risk fraction, economic growth
Introduction
One of the major problems that most modern countries face today is organized crime;’ its effects on
a society are pervasive. They allow drugs to be sold in schools and seduce the poor with easier
illegal incomes. The inside battles for power in this underground society and its continuous fight
with the authorities are amplified by the media decreasing the country’s level of perceived safety
for investors.
Specifically, organized crime greatly slows down the economic growth of a country. In countries
characterized by high corruption, a large portion of public investment will be consumed by criminal
organizations, resulting in deficient public services and infrastructure which will lower the general
competitiveness of the country and will fail in creating the expected employment. Crime
organizations such as the “Mafia” will finance their illegal activities (drugs, weapons, and alcohol)
by requiring “protection” payments from private businesses. They will also launder their illegal
incomes by starting private businesses characterized by high capital intensity (trucking,
construction, etc.) which will not operate according to the same rules as their competitors (no
financial constraints, profitability is not required, etc.) and will hence lower the profits of those
' Organized crime has had different definitions depending on the time and socio-economic area in which the
phenomenon has been analysed. While the concepts that will be developed in this paper do not require one clear
industries. Organized crime will lower the propensity of existing companies to reinvest their profits
in the same sector (if not force them to leave) and will create an eroding goal situation where local
firms will find it more convenient to stay small. Also, organized crime significantly hinders the
ability of a country to attract intemational investments, further decreasing employment
opportunities [1].
Among these effects, a minimum common denominator may he identified in the expected future
safety of the investment environment. There is a clear link between the presence of criminal
organizations in a country, the safety expected by the business community and the amount of
investments made in the country. This relationship is well known to public policy makers;
however, in many countries policies implemented to defeat criminal organizations have shown little
effect and the low investment level constitutes a serious obstacle to their economic development
[2].
This paper will suggest the structural factors that allow organized crime to survive and even prosper
in certain societies and, through hypotheses that will need to be further discussed and validated, will
show the potential of system dynamics models to support policy makers in setting objectives and
strategies to combat organized crime.
definition, a large literature has been developed on this issue [see Schelling, T., 1984 Choice and consequence.
Perspectives of an Errant Economist. Harvard University Press, Cambridge, (Ma) at all.].
2
Developing a causal loop diagram
An intuitive diagram
With a great deal of simplification, the traditional mental model that drives the decisions made in
the public safety sector is summarized in the causal loop diagram showed in Figure 1:
salaries
= savi Reg
taxes (R4A C
private
investments
pub inv in the priv sector
mae 4 unemployed
eee T to decide to stop crime
inv jn law enforcement inv i i inv i i
inv in Justice inv in Education as
d
rceived investment(safety pRB)
General level of education
Law enforcement effectiveness p> T to convict criminals
% Oe
y —~—_ S criminals imprisoned
risk fm crime
La
oink
FC unetmpl tempted by crime
in required from criminal activities
Nees
evel of corrubtion
crime attractiveness
Figure 1 — An intuitive causal loop diagram
The diagram clearly reflects a mental model dominated by reinforcing loops, which will now be
described in greater detail.
There are three reinforcing loops driven by increasing private investments made in a country:
R1) more investments => less unemployed => less unemployed tempted by crime => less
criminals => higher perceived investment safety => more investments
R2) more investments => less unemployed => higher salaries => more savings => more
investments
R3) more investments => less unemployed => higher salaries => less crime attractiveness =>
less criminals => higher safety perceived => more investments
Possible policies against organized crime also have reinforcing effects on the above loops, however
they constitute reinforcing loops themselves as described below:
Government investments in the private sector:
Government investments in the private sector represent public money allocations aimed to finance
new investments, mainly in infrastructures. Besides the mid term benefits deriving from better
infrastructure in a country (e.g. higher business efficiency and hence higher attractiveness for new
businesses), these initiatives determine a sudden creation of new jobs, hence so reinforcing the R1,
R2 and R3 loops, and generating the R4 reinforcing loop:
R4) more public income invested in the private sector => less unemployment => higher salaries
=> higher income from taxes => higher public income => more public income invested in
the private sector
Investments in the law enforcement system:
Investments in the law enforcement system traditionally consist of higher salaries, higher hiring
rate, better training, more research and development of new technologies aimed at increasing the
effectiveness of investigations. The reinforcing loops showed in the diagram are:
R5) more public income invested in the law enforcement system => higher police effectiveness
=> higher risk perceived from criminal activity => higher gain required from criminal
activities => lower crime attractiveness => lower fraction of unemployed tempted by crime
4
=> less criminals => higher perceived investment safety => more investments => less
unemployed => higher salaries => higher income from taxes => higher public income =>
more public income invested in the law enforcement system
R6) more public income invested in the law enforcement system => higher law enforcement
effectiveness => lower level of corruption => lower crime attractiveness => lower fraction
of unemployed tempted by crime => less criminals => higher perceived investment safety
=> more investments => less unemployed => higher salaries => higher income from taxes
=> higher public income => more public income invested in the law enforcement system
R7) more public income invested in the law enforcement system => higher police effectiveness
=> more criminals imprisoned => less criminals => higher perceived investment safety =>
more investments => less unemployed => higher salaries => higher income from taxes =>
higher public income => more public income invested in the law enforcement system
Investments in the justice system:
Investments in the justice system consist of all kinds of investments (both in human resources and
new technologies) aimed to increase the justice system’s efficiency. While this efficiency could be
viewed both in terms of ability to identify the guilty and in terms of higher speed, only the latter
will be taken into consideration.
The intuitive reinforcing loop generated by a more efficient justice system is:
R8) more public income invested in the justice system => less time required to convict criminals
I
Vv
=> more criminals imprisoned => less criminals => higher perceived investment safety
I
Vv
more investments => less unemployed => higher salaries => higher income from taxes
higher public income => more public income invested in the justice system
Investments in the education system:
Are represented by annual public expenditures such as higher salaries for public school teachers,
more scholarships, lower university taxes and alternative policies to increase the level of elementary
education in a country (e.g. financial incentives to poorer families to encourage them to send their
children to school, which could be determined by the children’s grades). A fraction of these
expenditures could also be invested in technologies, libraries, etc.
The reinforcing loop generated by such investments has been described as:
R9) more public income invested in the education system => higher general level of education
=> lower fraction of unemployed tempted by crime => less criminals => higher perceived
investment safety => more investments => less unemployed => higher salaries => higher
income from taxes => higher public income => more public income invested in the
education system
A more complex and dynamic view
The following considerations will explain the necessity of a more thoughtful description of the
system, explain why a more complete evaluation and quantification of the effects that a policy has
on a society needs to go through the concept of delays and feedback loop dominance, thus justifying
the introduction of a system dynamics methodology.
* A causal loop understanding of this complex system could be useless or even misleading if
it is not followed by an accurate quantitative analysis; in fact, the number of reinforcing
loops generated by each policy does not necessarily reflect its strength and effectiveness.
¢ In order to better evaluate the consequences of a certain policy on the system it is
necessary to search for eventual balancing loops and take them into proper consideration.
Some are fairly easy to identify and are often considered unavoidable side effects (e.g.,
every society has a certain level of corruption, a clear obstacle to the efficient allocation of
public finances), while others are less clear, sometimes counterintuitive, and often linked
to along term perspective.
¢ The above mentioned policies do not necessarily show their effects right after their
implementation, and these effects may have a different short, medium and long term
strength and intensity.
Consistent with the above considerations, the remaining part of this paragraph will describe some
“missing” loops that have been found and implemented in a system dynamics model. Figure 2 is a
direct consequence of the rent-seeking strategies that often characterize criminal organizations [5].
Figures 3a and 3b show the long-term effects of policies based on repression and punishment, and
portray the partial inability of such policies to determine structural changes in a society. Figure 4
shows one long term effect of higher investments in education (B6), but also describes the
reinforcing loops determined by the effects on what has been defined embeddedness [6], as it will
be better explained in other parts of this paper (see page 16).
6
private
+” investments
level of corruption
+
illegal opportunities
perceived investment safety
pot ind income fry crime
RB2/
crime attractiveness
criminals wy + a?
“ FC unempl tempted by crime
Rp)
crimes per criminal
Figure 2 — Consequences of the rent-seeking organized crime strategies
risk fm crime
investment in /" +
law enforcement pm
Law enforcement
minimum gain required
effectiveness
from criminal activities
public income hB4/ : AB5/
criminals imprisoned
taxes investments
oN .
unemployed
perceived ee safety FC unetmpl tempted by crime
Ne
Figure 3a - Long term effects in the law enforcement system
ee
crime attractiveness
salaries
criminals imprisoned . 2
criminals unemployed
af
FC unetnpl tempted by crime
Figure 3b — Long term effects in the justice system
public income
inv in Education
salaries
general Edu level
nvestments
+
FC of criminals willing to stop
safety perceived
¥
criminals ~q—___—~corrruption’
, wR?
=C unemp! tempted by crim
unemployed
Figure 4 — Effects of investments in the education system on the level of organized
crime’s embeddedness in a society
As the diagrams show, each policy has a different impact on “crime attractiveness” and generates
balancing loops in the system through what has been defined “Fraction of unemployed tempted by
crime” and “Fraction of criminals willing to stop”. Both categories constitute what will be defined
in the rest of this paper as the “risk fraction”, whose conceptualization becomes central for a correct
evaluation and implementation of the feedback loop diagram in a system dynamics model and the
formulation of its dynamic hypotheses.
The following pages will focus on this concept and on its implementation in a system dynamics
model.
A dynamic hypothesis: the risk fraction
The concept of the risk fraction originates from the long ago observed existence of an underworld
and an upperworld in almost every society [3]. In system dynamics terms it could be inferred that,
in most societies, there is a continuous flow between unemployed and criminals, as well as there are
irreducible criminals and there are unemployed whose ethics will prevent them to commit illegal
actions.
What is left is therefore a fraction of unemployed and a fraction of criminals, people that are
currently unemployed and can choose between searching for employment and what has been
defined as the criminal option [4]. This fraction constitutes the recruitment base for crime
organizations, and can determine their strength and presence in a given society.
As such, the risk fraction can greatly influence the safety perceived by potential investors and the
potential growth of a country, hence assuming a central role.
Furthermore, in a simplified setting where there is no direct flow between workforce and criminals,
there is no in and out migration, and no other forms of criminality are taken into consideration, the
dynamics of this risk fraction set the basis for the dynamic hypothesis of the model.
Using a system dynamics perspective, the fraction of unemployed tempted to commit criminal
actions is influenced by variables such as the unemployment rate, the average cultural level of the
society, the availability of opportunities to commit crime (which depends on the level of corruption
of the society and the economic growth of the country). The size of this fraction as well as the time
that it will take to actually become an inflow into the stock of criminals depend on the attractiveness
of the potential income deriving from illegal actions compared to other income opportunities and
discounted by the gain expected from crime which depends on the risk perceived. Similar thoughts
can be formulated for the fraction of people abandoning the criminal lifestyle.
Is it possible to identify and quantify this fraction in a society? If yes, how homogeneous is it? In
other words, would it be correct to generalize the causes that determine its size and determine its
changes in a society?
System dynamics cannot answer these questions. But while the validity of the hypotheses
developed in this paper greatly depends on these answers, a system dynamics approach can greatly
contribute to a more profound understanding of its dynamics and of its socio-economic
consequences. Lets assume that it is indeed possible to identify this fraction and generalize its
dynamics: the next paragraph will then describe one way in which this fraction could be “scanned”
using a system dynamics perspective.
Defining crime attractiveness and modeling the risk fraction
Figure 5 shows the role and meaning of “crime attractiveness” as is intended in this work.
4
current_average_sala a p bs id€S_ unemployed
NA
new_industrial_investments|
criminals average_ilegall_income
,
t
\
initial_gain_from_criminal_activities
nA
delayed_effect_of_police_effectiveness_on_exp_gain
Figure 5 — The crime attractiveness sector of the system dynamics model
The main variables are:
Crime attractiveness
units dimensionless
10
aux crime attractiveness = discounted average criminal income/average income option
doc crime attractiveness depends on the potential income deriving from illegal activities
discounted by the risk of crime and compared to a weighted average of the average legal salary and
the unemployment subsidies.
Criminal income option
Units $ per person per month
aux criminal income option = average potential illegal income/required gain from
criminal activities
doc criminal income option represents the potential individual illegal income
discounted by the perceived risk deriving from criminal activities.
Gain required from criminal activities
Units dimensionless
aux gain required from criminal activities = delayed effect of law enforcement effectiveness on
the gain required from criminal activity * reference gain required from criminal activities
doc gain required from criminal activities is the gain required from criminal activities represents
the multiple of average legal income required to compensate for the risk involved in the crime,
causing one to choose criminal activity over legal employment. It can be influenced by the law
enforcement effectiveness. It is assumed that stronger penalties will not have a relevant influence on
the risk perceived and hence on the gain required.
Average illegal income
Units $ per person per month
aux average illegal income = potential illegal incomes / (1+criminals)
doc average illegal income is the fraction of the potential illegal incomes divided by the number
of criminals determines the individual average illegal income of the average criminal enrolled in a
criminal organization. Under extreme conditions, if there are no criminals we must assume that the
entire fraction represents the potential income and can therefore attract new criminals into the
system.
Potential illegal income
Units $ per month
aux potential illegal incomes = new industrial investments * corruption
11
doc potential illegal incomes represent the fraction of new investments that could potentially
become a source of illegal revenues for criminal organizations. The higher the corruption is in a
country, the higher this fraction will be. No corruption means that no fraction of the capital
accumulated in a region can be “taken away" by criminal organizations. Many criminal
organizations drain money from private businesses to finance their illegal activities. In this
microworld, no corruption would also mean no more drainage, hence no financial resources, no
profitability for organized crime and, ultimately, no organized crime.
Legal income option
Units $ per month per person
aux legal income option = (unemployed * unemployment subsidies + current average salaries *
workforce) / Total adult population
doc legal income option represents a synthetic indicator of the potential salary available to an
unemployed as an alternative option to an illegal “career”. It is calculated as a weighted average of
the unemployment benefits and the salary available with a normal job: the unemployment rate
represents the probability (hence the weight) to earn either one.
Effects of crime attractiveness on the risk fraction
a) Effect on unemployed
aux — effect on unemployed = GRAPH
doc effect on unemployed is the effect of crime attractiveness on the fraction of unemployed
tempted by crime as described by the graph below. When attractiveness is 1, a 5 % of the total
unemployed will be tempted by crime. This fraction will sensibly grow as the attractiveness
increases above its reference value of 1 and will be less sensitive to attractiveness changes below 1.
Under extreme crime attractiveness values, it is assumed that a maximum 20% of unemployed will
consider the possibility of joining organized crime; on the other hand a minimum 3 % approx. will
continue to be tempted by organized crime even when its attractiveness is very low (in other words,
it is assumed that there will always be a small percentage of unemployed that for various reasons
will join organized crime regardless the fact that the crime option is not attractive: people generally
are not as rational, and one can only model the way "most" of the people behave under normal
circumstances).
12
Jutput [effect_on_unemployed]
0,20-
O6 2.0
b) Effect on criminals
aux effect on criminals = GRAPH
doc — effect on criminals originates from the hypothesis that, at any moment in time, there is a
certain fraction of criminals who, for various reasons, chose to stop committing crime and will enter
the stock of unemployed. The way this fraction responds to changes of crime attractiveness is
described by the curve below; when attractiveness of crime is 1, 5 % of the criminals is assumed to
stop crime; this percentage increases and saturates at approximately 13 % when attractiveness is
very low, meaning that crime attractiveness has a weak influence on the percentage of criminals
choosing to stop crime. This percentage will decrease to a minimum of approximately 1 % if crime
attractiveness reaches very high values.
Jutput (effect_on_criminals)
0.20-
13
As stated in other parts of this paper, lack of empirical evidence does not allow to verify the
hypotheses behind the described effects. The main purpose is to evaluate their effects on the system
and to create a basis for further discussions and research in the field.
Figure 6 below gives an overview of how the “risk fraction” has been represented.
eff_of_poll
ae
@
eff_of_low_unem|
effeq
lel_eff_fm_inv_in_edu_on_eff_of_crime_attractiveness_on_unemp
orkforce
ment
ref_justice_speed
b
NA
Kit of \_in_justice_efficiency_.
ih, all
efff of_edu_on_time ©
t_to_change ref_time
A
Ip|_on_new_hiring
A
L_on_unemployed
fc_of_unempl
ime
imprisonment_time
ent_ory
cE.
d_emkloyme
T_to_fit
loye
| further_attrit
unemployed
u
Figure 6 — Modeling the risk fraction
14
Z|
del_eff_fm_inv_in_edu_on_eff_of_attractiveness_on_crin
Alternative measures of policies effectiveness
Using a system dynamics perspective and according to the definition of crime attractiveness and
risk fraction that has been described, the effectiveness of the various policies, hence their ability to
deplete the stock of criminals and allow the economic system to grow, depends on how they
influence these variables over time.
The causal loop diagram previously developed describes some of the effects that additional
investments in the education, justice and law enforcement have on the entire system. The effects
that have been taken into consideration are:
Effect of investments in education on corruption
Effect of investments in education on time to become criminal
Effects of investments in education on the risk fraction
Effect of investments in law enforcement on corruption
Effect of investments in law enforcement on gain required from illegal activities
Effect of investments in law enforcement on criminals imprisoned
Effect of investments in the justice system on its speed
Each one of these effects raises questions and concerns regarding their estimation and
quantification. Those effects that have been found most pertinent and may stimulate interesting
discussion will now be described in detail.
1) Effect of education on corruption
The assumption underlying the curve below is that a policy aimed at increasing the level of general
education will create a better society where corruption will be much less accepted and business
owners will generally be more reluctant to comply with requests for money from organized crime
organizations. The curve describing these effects is a mere hypothesis. It implies that, regardless of
the efforts concentrating on an education policy, 40% of corruption can be eliminated in this
fashion. It can be reasonably inferred that a policy focused on higher general education will start to
influence the level of corruption of a society only when younger generations will become part of the
workforce. This delayed effect is simulated through a third order information delay with a ten-year
adjustment time.
15
IN
2) Effect of law enforcement on corruption
The function relies on the assumption that higher technologies available to the police can
significantly help reduce the level of corruption. Initial additional efforts result in remarkable
decreases in the corruption level, but as corruption decreases it becomes more and more difficult to
eliminate the remaining layers of corruption. The model assumes that only 50% of corruption can
be eliminated through police investigations.
16
3) Effects of education on the risk fraction.
As previously shown (see R11 and R12 loops at page 8) a high level of investments in education in
a society has positive long term effects on its shared values, and contributes to decrease the level of
embeddedness of criminal organizations. In fact, the embeddedness does not represent here those
values able to promote the competition and co-operation which set the basis of the social structure
of a market, hence becoming a basic ingredient in the process of development of any society. In
this context it indicates a perverse mechanism that almost makes a criminal option appear socially
acceptable and that promotes rent-seeking behaviors, legitimizing protection, violence and even
certain forms of racketeering [7]. The more organized crime is embedded in a society, the less
traditional policies to contrast crime will prove to be effective.
a) Effect of education on unemployed
The assumption in the model is that, ceteris paribus, the higher the cultural level of a country the
lower the crime attractiveness will be. It is assumed that 50% of the fraction of unemployed
tempted by crime can be prevented from deciding to actually become criminal when maximum
effort is put on this policy. Initial efforts will results in encouraging results, but as the policy is
implemented increasing marginal efforts need to be put in order to achieve remarkable results: we
get close to the fraction of unemployed tempted by crime that cannot be reached or influenced by
this policy.
ut (eff_of_|_in_education_on_crime_tendeneyl]
1.0
= SF
a 1
=
a
10
The delay between the implementation of such a policy and its impact on society has been modeled
as a third order, with a five-year adjustment time.
17
b) Effects of education on criminals
Investments in the education system can also influence the fraction of criminals that could stop
crime. Here the effect is more subtle, and is also less remarkable that the effect on unemployed. It
is easier to prevent an unemployed person from starting crime than to convince a criminal to
abandon crime. In this respect, investments in the education system are not very efficient. Higher
investments in this sector will probably not change the ethical values of a criminal. A general
increase in the level of culture and education, in other words a higher civil awareness in the poorest
classes of our society, could nonetheless influence the decision of a criminal to stop crime when
less opportunities are available. It takes a remarkable financial effort for an extended period of time
to achieve meaningful results. The curve is hence sensitive to high levels of financial resources
invested in this policy; it is assumed that a maximum 20% of the fraction can successfully be forced
out of it through this policy. It is hence assigned a maximum value of 1,2 to this curve.
Output (eff_of_education_on_decision_to_stop_crime)
10
Building civil awareness is a process that takes generations, therefore this effect has been modeled
as a third order information delay with a 25-year adjustment time.
Policy evaluation
Given the above described assumptions and hypotheses, different policies have been evaluated
through their impact on the microworld that has been portrayed. The outcome of the simulations
that will be shown very much depends on the shape of the curves, however some indications and
learning can be inferred also at this early stage of development.
The results have been evaluated in the model through an “instant welfare indicator’, a synthetic
socio-economic score-keeping indicator which grows as the economy grows and decreases as the
level of crime increases. It has been calculated as follows:
18
(industrial investments + investments from savings + consumption)/relative number of crimes,
where
relative number of crimes = number of crimes / reference number of crimes.
Time is measured in months and when each simulation starts the instant welfare indicator equals
158.
Simulation 1 portrays the base behavior, where no additional investment in either sector is made.
The model endogenously generates the financial resources available for each policy; they result
from the taxation of the average salaries earned by the workforce. It is assumed that each policy
utilizes 100% of the available resources.
Figure 7 shows the changes of the instant welfare indicator over an 800 month (67 years
approximately) time horizon as a consequence of each policy implemented, while Table 1
summarizes the results.
The policies have been indicated with the following indications:
I: additional investments in Infrastructure
L additional investments in the Law enforcement system
E: additional investments in the Education system
J additional investments in the Justice system
Sim. | Policies mixes over time % I1- E- L-J Score
Time 0 Time 200 Time 400 Time 600
1 0-0-0-0 0-0-0-0 0-0-0-0 0-0-0-0 190
2 1-0-0-0 1-0-0-0 1-0-0-0 1-0-0-0 176
3 0-1-0-0 0-1-0-0 0-1-0-0 0-1-0-0 199
4 0-0-1-0 0-0-1-0 0-0-1-0 0-0-1-0 200
5 0-0-0-1 0-0-0-1 0-0-0-1 0-0-0-1 187
6 0-0,5-0,3-0,2 |0-0,5-0,3- 0,2 0-0,5-0,3-0,2 |0-0,5-0,3-0,2 |576
7 0-0,7-0,2- 0,1 0,1- 0,6-0,2-0,1 [0,2-0,4-0,3-0,1 |03-0,3-0,3-0,1 |724
Table 1 — Policies specification
19
t t t
200 400 600
Time
Figure 7 — Welfare indicator responses to policies
The results showed in figure 7 suggest that in the long run, after a transient behavior due to the
initial conditions, the system finds its own equilibrium. In addition to this, the other runs suggest
that various policies most likely have a destabilizing effect on the system or, as in Simulation 3, will
stabilize it at a slightly higher level. Finally, Simulation 7 suggests that the most effective results
are products of a profound understanding of the dynamics of the environment in which they are
implemented and by the ability of policy makers to correctly change their strategies over time.
20
Conclusions
This paper suggests that system dynamics modeling has much to offer the field of political economy
and that it can be a powerful tool to help policy makers understand the role of feedback in complex
dynamic systems, quantify their effects in a social environment and choose the most effective
policies over time. Each effect described in the paper underlines hypotheses that need to be
validated and that raise questions which have not been answered or even sufficiently explored.
From this respect, this paper has ventured to prove once again the vast potential for system
dynamics methods to communicate ideas, compare views, and stir thinking.
References
[1] Centorrino, M. - Signorino, G. 1997
Macroeconomia della mafia, Nis, Roma.
[2] Rodano, G., Fornari, L., Giacomelli, S. 2000
Gli effetti economici delle attivita criminali e le politiche di contrasto, Uic, Roma.
[3] Kefauver, E. 1953
Il gangsterismo in America, Einaudi, Torino (original title: Crime in America).
[4] Becchi, A. 2000
Criminalita organizzata, Donzelli, Roma.
[5] Tollison, R.D. 1995
The Economic Analysis of Rent Seeking, Elgar, Aldershot.
[6] Granovetter, M. 1989
Azione economica e struttura sociale: il problema dell’embeddedness, in Azione economica
come azione sociale, Franco Angeli, Milano.
[7] Ben-Porath, Y. 1980
The F-connection: Families, Friends and Firms in the Organization of Exchange, in
Population and Development Review, 6. Bevilacqua, P. 1998
21
