DEPÓSITO LEGAL ZU2020000153  
Esta publicación científica en formato digital  
es continuidad de la revista impresa  
ISSN 0041-8811  
E-ISSN 2665-0428  
Revista  
de la  
Universidad  
del Zulia  
Fundada en 1947  
por el Dr. Jesús Enrique Lossada  
Ciencias  
Sociales  
y Arte  
Año 12 N° 34  
Septiembre - Diciembre 2021  
Tercera Época  
Maracaibo-Venezuela  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Trafficking in persons: cluster analysis of sending and receiving  
countries  
Yulia A. Siluyanova *  
ABSTRACT  
The objective of this article is to deepen the understanding of the problem of human  
trafficking by analyzing the patterns of the distribution of traffic across countries, based on  
international statistics on various socio-economic indicators. We conducted cluster analysis  
using neuro network grouping algorithm of Kohonen self-organizing maps, basing on 44  
variables reflecting different social and economical aspects for 144 countries. Countries were  
grouped according to the type and level of risk of trafficking-related crimes, and traffic  
distribution maps were built based on generally accepted hypotheses about traffic factors.  
As a result of the study, a number of hypotheses explaining the nature of traffic were tested.  
The results reveal the linkage between the risk of incoming and outgoing trafficking and the  
socio-economic parameters of the countries and groups.  
KEY WORDS: human trafficking; slavery; exploitation; trafficking in persons; international  
crime; cluster analysis; Kohonen self-organizing maps; distribution maps; migration; Human  
Rights.  
*PhD student, Lomonosov Moscow State University Moscow, Russia. ORCID:  
https://orcid.org/0000-0001-9752-0472. E-mail: zernovaju@gmail.com  
Recibido: 14/06/2021  
Aceptado: 05/08/2021  
317  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Trata de personas: análisis por grupos temáticos de los países de  
origen y de acogida  
RESUMEN  
El objetivo de este artículo es profundizar en la comprensión del problema de la trata de  
personas mediante el análisis de los patrones de distribución del tráfico entre países, sobre la  
base de estadísticas internacionales sobre diversos indicadores socioeconómicos. Se realizó  
un análisis de conglomerados utilizando el algoritmo de agrupación de redes neurológicas de  
los mapas autoorganizados de Kohonen, basándose en 44 variables que reflejan diferentes  
aspectos sociales y económicos para 144 países. Los países se agruparon según el tipo y el  
nivel de riesgo de delitos relacionados con la trata, y se construyeron mapas de distribución  
del tráfico basados en hipótesis generalmente aceptadas sobre los factores del tráfico. Como  
resultado del estudio, se probaron una serie de hipótesis que explicaban la naturaleza del  
tráfico. Los resultados revelan el vínculo entre el riesgo de tráfico entrante y saliente y los  
parámetros socioeconómicos de los países y grupos.  
PALABRAS CLAVE: trata de personas, esclavitud, explotación, trata de personas,  
delincuencia internacional, análisis de conglomerados, mapas autoorganizados Kohonen,  
mapas de distribución, migración, Derechos Humanos.  
Introduction  
According to the estimates of international organizations, the number of victims of  
human trafficking around the world exceeds 40 million people at the moment, and the figure  
is steadily growing every year.  
Trafficking, merging into migration flows, becomes almost invisible, which increases  
its latency, reduces risks for attackers and ensures the smooth operation of the criminal  
conveyor” of the human trafficking industry. In the context of globalization, which has made  
the population more mobile and national borders more permeable, the profits of criminal  
structures from the sale and exploitation of “human goods” amount to more than 150 billion  
dollars a year (ILO says forced labour generates annual profits of US$ 150 billion (ILO says  
forced labour generates annual profits of US$ 150 billion, 2014).  
This figure is comparable to the budgets of countries such as Finland and Saudi  
Arabia. At the same time, the most vulnerable in the face of the threat are representatives of  
the most unprotected social groups: orphans, the poor, people from disadvantaged families,  
the disabled, women, children, the elderly.  
318  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Despite the attempts of the international scientific community to study the origins of  
the problem, the creation of international analytical institutions dedicated to the study of the  
phenomenon of trafficking in persons, the number of registered cases of this kind of crime is  
growing every year. This may mean that the number of victims is indeed increasing, and the  
ability to identify them is growing in the world.  
Researchers are faced with the task of systematizing the available data on traffic,  
generalizing them and building such analytical models. Reliable assessment and analytics of  
trafficking processes are key to validating the theories on which the scientific community  
relies today. Having studied the phenomenon of trafficking, it is possible to reach a new level  
of understanding of the problem of trafficking in persons and formulate useful  
recommendations for combating trafficking in persons for state and non-state structures.  
Due to the descriptive approach to the problem, there is still no generally accepted  
established explanation for this phenomenon. An analysis of scientific literature has shown  
that ideas about its nature are very different.  
It should be noted that the problem of trafficking in the scientific community is  
relatively new and complex theories that fully reflect its causes, preconditions and patterns  
have not been developed to date. Trafficking is often considered within the framework of one  
of the scientific disciplines, such as law and forensics. Moreover, the phenomenon itself  
requires an interdisciplinary study, since it has deep social, cultural and anthropological  
roots.  
Many researchers from the scientific community are trying to explain and predict the  
dynamics of the development of these processes. The same task is facing the governing bodies  
of states and international organizations. A serious problem for researchers is the lack of  
reliable empirical data and analytical models based on such data and allowing them to make  
validated informed assumptions about the patterns of trafficking.  
In this paper, we attempt to deepen our understanding of the problem by examining  
the patterns of distribution of countries by the level and nature of trafficking. By grouping  
countries with similar trafficking crime situations, we can assess socioeconomic markers  
that indicate risk groups.  
The problem of the lack of orderly reliable sources describing the situation with  
human trafficking is still urgent. Researchers and international experts are forced to rely on  
319  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
estimates of varying degrees of reliability in their work. The empirical data used by  
researchers is either fragmentary and has many gaps, or is judgmental, subjective.  
Thus, we can use existing comparable databases with low confidence, or relatively  
reliable empirical data that are too narrow and local to draw general conclusions and not  
comparable with each other.  
The objective of this article is to deepen the understanding of the problem of human  
trafficking by analyzing the patterns of the distribution of traffic across countries, based on  
international statistics on various socio-economic indicators. We have concentrated on  
those data, information about which is homogeneous, comparable and comes from reliable  
sources. In our case, these are data from the UN and other international organizations, or  
large transnational NGOs.  
1
. Methodological aspects  
We included 44 variables in the study, reflecting various social, economic,  
demographic, cultural and other indicators. Here is a complete list with explanations.  
1
.1. Trafficking and slavery  
Global Slavery Index (slavesТotal). The absolute number of victims in exploitation,  
according to the Global Slavery Index for 2018 (Global Slavery Index 2018, 2018).  
Population in slavery per 1000 (SlavesPer1000). The number of victims of exploitation  
per 1000 population, according to the Global Slavery Index for 2018 (Global Slavery Index  
2018, 2018).  
Country of exploitation (inExpl). Ranking estimate of the volume of exploitation  
according to the CTDC portal of the International Organization for Migration, 2018. The  
indicator is ranged from 1 to 5, where 1 corresponds to the flow of less than 10 victims, 2 -  
from 10 to 220, 3 - from 220 to 930, 4 - from 930 to 5600, 5 - more than 5600 (International  
Organization for Migration, 2019). This variable reflects the level of exploitation in the  
country and includes the incoming traffic flow; in the course of the study, we will consider it  
as an exploitation factor.  
Country of origin (sourceCountry). Ranking of outbound traffic according to the  
CTDC portal of the International Organization for Migration, 2018. The indicator is ranged  
from 1 to 5, where 1 corresponds to the flow of less than 10 victims, 2 - from 10 to 220, 3 - from  
320  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
220 to 930, 4 - from 930 to 5600, 5 - more than 5600 (International Organization for  
Migration, 2019). This variable reflects the origin of the victims of trafficking identified in  
different countries.  
K trafficking. Generalized traffic ratio based on statistics from the CTDC portal of the  
International Organization for Migration. Calculated as the sum of the squares of the  
previous two variables.  
Google HT. The volume of mentions in the Google news aggregator for the search  
query “human trafficking” per year.  
1
.2. Demography  
Population. The absolute indicator of the number of population in the country  
according to the UN data for 2016 (Population data, 2019).  
Birth rate. Birth rate - per 1000 people (Birth rate, crude (per 1,000 people), 2019).  
Fertility rate. Special fertility rate (per woman) (Fertility rate, total (births per  
woman), 2019).  
Adolescent fertility rate. Adolescent fertility is the number of births per 1000 women  
aged 15-19 years (Adolescent fertility rate (births per 1,000 women ages 15-19), 2019).  
Death rate. Mortality rate per 1000 people (Death rate, crude (per 1,000 people), 2019).  
Life expectancy at birth. Life expectancy at birth (Life expectancy at birth, total  
(
years), 2019).  
Population growth. Population growth for the year in percent (Population growth  
annual %), 2019).  
Population ages 65 and above. The proportion of the country's population aged 65 and  
over, expressed as a percentage of the total population (Population ages 65 and above, total,  
019).  
(
2
Population ages 0-14. The proportion of the country's population between the ages of  
and 14, expressed as a percentage of the total population (Population ages 0-14 (% of total  
population), 2019).  
0
1
.3. Migration  
International migrant stock. The number of international migrants in the country as a  
percentage of the total population of the state (International migrant stock (% of  
321  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
population), 2019). This indicator is calculated on the basis of population censuses and  
includes people living in the country, but born in another state. In the absence of data, the  
UN independently estimates the indicator.  
Refugees Azyl. The absolute number of refugees who arrived in the country (Refugee  
population by country or territory of asylum, 2019).  
Refugees Orig. The absolute number of refugees who left the country (Refugee  
population by country or territory of origin, 2019).  
Status Azyl pending. Number of pending applications for asylum or refugee status  
(
Mid-Year Trends 2017, 2017).  
Migrant remittance inflow. Total amount of migrant remittances in millions of US  
dollars from other countries during the year.  
Migrant remittance outflow. Total amount in millions of US dollars sent by migrants  
via remittances to other countries (Migration and Remittances Data, 2019).  
Mobile cellular subscriptions. The number of new mobile connections per 100 people  
per year (Mobile cellular subscriptions (per 100 people), 2019).  
1
.4. Economy  
Adjusted net national income. Net national income per capita in US dollars (NPI)  
(
(
(
Adjusted net national income per capita (current US$, 2019).  
Unemployment. Unemployment as a percentage of the total labor force  
Unemployment, total (% of total labor force) (modeled ILO estimate), 2019).  
Employment 15-24. Employment among young people aged 15 to 24 years  
Employment to population ratio, ages 15-24, total (%)(modeled ILO estimate), 2019).  
Inflation. Inflation is the GDP deflator index in percent per year (Inflation, GDP  
deflator (annual %), 2019).  
Natural resources rents. The share of revenues received by the government from  
natural resources, expressed as a percentage of GDP (Total natural resources rents (% of  
GDP), 2019).  
1
.5. Inequality  
GINI index. Gini coefficient, reflecting economic inequality in society (GINI index  
(
World Bank estimate), 2019).  
322  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Gender Inequality Index. Gender Inequality Index. This indicator was introduced by  
the UN to quantify gender discrimination (Gender Inequality Index, 2019).  
Women in parliament. The proportion of seats in the country's parliament occupied  
by women (Proportion of seats held by women in national parliaments (%), 2019).  
1
.6. Conflicts  
Global Peace index. The Global Peace Index reflects the level of security in a country  
or region, taking into account different types of crime, internal and external conflicts,  
terrorist threats, arms sales, etc. The countries with the lowest index are the safest to live in  
(
Global Peace Index, 2019).  
Battle-related deaths. The number of deaths during armed conflicts (Battle-related  
deaths (number of people), 2019).  
1
.7. Education, healthcare, ecology, culture and other  
Corruption. Corruption perception index, which reflects the abuse of authority by  
civil servants. It is measured on a scale from 100 (no corruption) to 0 (very high level of  
corruption) (Corruption Perceptions Index, 2019).  
Gov expenditure on education. Government spending on education, expressed as a  
percentage of GDP (Government expenditure on education, total (% of GDP), 2019).  
Literacy. Literacy as a percentage of the total population (Literacy rate, 2019).  
Secondary education. Population over 25 years old with at least secondary education,  
as a percentage of the total population (Key charts on Education, 2019).  
Tuberculosis. The incidence of tuberculosis per 100,000 people (Incidence of  
tuberculosis (per 100,000 people), 2019).  
Undernourishment. Percentage of the population at risk of hunger (Prevalence of  
undernourishment (% of population), 2019).  
Rural population. The rural population is the percentage of the country's population  
living in rural areas (Rural population (% of total population), 2019).  
Forest area. Forest area in the country (List of countries by forest area, 2019).  
UNESCO. The number of UNESCO heritage sites in the country (UNESCO World  
Heritage Centre, 2019).  
323  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Happiness Index. International index of happiness, which measures people's  
satisfaction with their living conditions, including income, prospects, environmental  
situation, etc. (World Happiness Report 2018, 2018).  
Freedom to make life choices. Freedom of choice. One of the variables included in the  
happiness index. This is the average of the responses to the question “Are you satisfied with  
the freedom to choose what you do with your life?”  
Social support. Social support (or having someone to count on in a difficult situation)  
is also one of the variables in the Happiness Index. This is the average of the binary responses  
(
0 or 1) to the question "If you are in trouble, do you have any family or friends you can count  
on?"  
In order to form a picture of the distribution of traffic in the modern world based on  
the listed data, we use cluster analysis. The algorithm that we use for classification in our  
model is self-organizing Kohonen maps. The result of his work is graphic maps, where each  
point represents one observation in the sample and has coordinates strictly defined by the  
network.  
2
. Results and discussion  
The analysis is supposed to undercover some important trends of trafficking flows,  
understanding that human trafficking is a complex phenomenon, having multiple economic  
and social causes (Siluyanova, 2019). In our work we will verify some suggestions about  
trafficking, offered by researchers. So, Lanier and Henry (Lanier, Henry 2004) argue that  
traffic is determined by the theory of rational choice conclude that the choice of victims is  
not determined by any certain laws, but is carried out situationally. Cameron and Newman,  
divide the factors influencing traffic into two groups: structural and direct (Cameron,  
Newman 2008). Louise Shelley identifies demographic parameters, such as population  
growth, as an important prerequisite for traffic, as well as the reduction of the rural  
population by moving people to cities or abroad (Shelley 2010). In addition, gender inequality  
and discrimination are important factors in shaping trafficking flows, as noted by Kara  
Siddarth (Siddharth, 2009). Alexis Arnowitz claims that all migratory flows, including  
human trafficking, have the same causes (Aronowitz 2009).  
We took the variables reflecting the absolute number of victims of exploitation from  
the Global Slavery Index database. We also have statistics on the number of registered cases  
324  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
of exploitation, which are kept by international institutions. As expected, these indicators  
should correlate with each other, since they are different measures of the same phenomenon.  
However, there is no relationship between these variables: the correlation coefficient of the  
inExpl and slavesPer1000 variables is -0.09. Since inExpl is an ordinal variable, we also  
checked Spearman's nonparametric correlation, however, no dependence was found: the  
coefficient was -0.08.  
The absence of any statistically significant correlation between the two datasets  
reflecting the level of trafficking indicates that, most likely, one of the variables does not  
correctly reflect the essence of the phenomenon under study, and we need to choose the one  
that is more representative. Since the statistics of actual registered cases recorded by  
international institutions seem to be more representative than the estimated value published  
by the NGO, we used the inExpl variable as a measure of the level of exploitation, and the  
sourceCountry variable to measure outgoing traffic.  
Already at this stage, we can be convinced that the empirical base hides in itself many  
contradictions and difficulties. This result of comparing data from international institutions  
and NGOs indicates a low degree of measurement of the problem. This result also confirms  
fears that NGO statistics are of unclear origin and distort the overall picture.  
First, let's build maps for the level of exploitation in the country (inExpl). To do this,  
we divide the data into five groups (Figure 1):  
С4  
С5  
С2  
С1  
С3  
Figure 1 - Map of clusters with grouping attribute inExpl  
325  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
In the figure, each point is a sample country. The stronger the countries are similar to  
each other in terms of a set of features, the closer they are displayed on the map. The points  
farthest from each other, respectively, have the maximum degree of difference.  
First of all, we consider the statistics of clusters (table 1) and give a description of each  
group. For convenience, all indicators in the table are highlighted in colors: green indicates  
the lowest values, orange indicates the highest values.  
Table 1 - Table of average values of clusters with the grouping attribute inExpl  
Cluster CountrExpl RemitOut  
NеtInc  
4662  
Corruption  
33,23  
Pop65+  
6,79  
Pop0-14 Fertility  
C 1  
1,896  
1,524  
2,108  
4
1015  
179  
4327  
11924  
15552  
27,22  
41,84  
16,6  
18,28  
22,19  
Int  
2,344  
4,638  
1,635  
1,75  
C 2  
C 3  
C 4  
C 5  
1156  
23635  
11532  
21408  
30  
63,14  
44,67  
47,25  
3,1  
17,05  
12,99  
2,97  
2,75  
2,277  
secEduc  
migrant  
stock  
2,47  
25+  
Literacy  
97,2  
Undernourish  
8,22  
Happiness  
5,353  
Rural pop  
Inflation  
3,35  
C 1  
61,9  
42,46  
60,59  
27,19  
33,86  
10,44  
C 2  
C 3  
C 4  
C 5  
33,8  
89,2  
92,9  
65,1  
72,8  
99,7  
99,5  
99,2  
23,72  
2,96  
2,32  
4,129  
6,246  
5,881  
6,128  
2,83  
11,1  
7,79  
3,9  
1,41  
4,7  
-13,08  
Life  
5,04  
54,65  
Death  
rate  
Mobile  
112,4  
Tuberculosis  
100  
GINI  
40,98  
42,26  
32,45  
33,79  
31,31  
GenInequality Unemployment  
exp  
C 1  
0,3985  
0,5788  
0,1218  
0,2163  
0,2836  
7,13  
7,72  
10,96  
4,74  
4,26  
6,42  
8,87  
9,76  
10,82  
2,83  
72,91  
61,04  
79,57  
74,4  
76,41  
C 2  
C 3  
C 4  
C 5  
77,8  
123,7  
126,2  
149,2  
238,2  
16  
52  
12,7  
Compiled by the author  
Cluster 1 (C1) - 33.3% of observations. This group of countries has a low per capita  
income, average inflation and a fairly high unemployment rate. The birth and death rates are  
average here, there are many children in the structure of the population, and the share of the  
rural population is quite high. There are few migrants in these countries, high morbidity,  
economic and gender inequality, and the level of corruption.  
Cluster 2 (C2) - 29% of observations. It has the lowest NPI, high unemployment, high  
inflation and the highest risk of hunger. These countries have record high mortality and  
326  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
fertility, many children, the lowest life expectancy, fewmigrants, the highest morbidity, very  
low literacy and education levels, strong corruption, the worst indicators of all variables  
reflecting the observance of human rights.  
Cluster 3 (C3) - 26% of observations. This is almost the complete opposite of the  
previous "dysfunctional" cluster: High NPI, low corruption, excellent indicators of the index  
of happiness and equality, low morbidity, low birth rate, high mortality, many migrants  
(
Siluyanova, 2019).  
Cluster 4 (C4) - 6% of observations. These countries show an average NPI, relatively  
high inflation rates, and lowunemployment. The birth rate here is low, and the mortality rate  
is very high. There are a lot of children and the elderly in the population structure, the average  
number of migrants, average indicators of corruption, happiness, inequality, high literacy and  
a high rate of remittances of migrants abroad.  
Cluster 5 (C5) - 6% of observations. Typical features are a high NPI, low  
unemployment, and no inflation. This is the average birth rate, very low mortality, high life  
expectancy, very few old people and a lot of migrants. At the same time, the indices of  
happiness and inequality are average, or high and a record indicator of outgoing remittances  
of migrants.  
We identified two “prosperous” groups, one group with dramatically poor  
performance, and two groups in between. The map we have built perfectly reproduces the  
geopolitical concept of dividing countries into developed, transitional and third world  
countries. One of these concepts - world-systems analysis - suggests dividing the world-  
system into core, periphery and semi-periphery with properties very similar to the identified  
properties of clusters. Thus, “safe” clusters (C5, C3) represent the core, “dysfunctional” (C2)  
the periphery, and the layer (C1) semi-periphery. We are even more convinced of the  
similarity of the classification with the world-system interpretation by examining the lists  
of clusters (Table 2).  
The prosperous clusters included the European states and Canada (C3) and the  
countries of the Middle East (C5). The disadvantaged cluster is the countries of Africa.  
Interlayer” (C1) - Asian and Latin American countries.  
Next comes the most important part of this stage of the analysis; this is an  
examination of the map of the distribution of individual variables (Figure 3).  
327  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Table 2 - List of countries included in clusters with grouping attribute inExpl  
С1  
С2  
С3  
С4  
С5  
United Arab  
Emirates  
Turkey  
Uganda  
Norway  
United States  
Russian  
Philippines  
Indonesia  
Senegal  
Haiti  
Macedonia  
Japan  
Federation  
Saudi Arabia  
Qatar  
Poland  
Moldova  
Malaysia  
Cambodia  
Uzbekistan  
Ghana  
Italy  
Lebanon  
Kuwait  
Afghanistan  
Czech Republic  
Bosnia and  
Thailand  
Tajikistan  
South Africa  
Egypt  
Zambia  
Yemen  
Herzegovina  
Kazakhstan  
Belarus  
Jordan  
Oman  
Albania  
United Kingdom  
Switzerland  
Sweden  
Sudan  
Romania  
Bulgaria  
Bahrain  
Sierra Leone  
Papua New Guinea  
Pakistan  
China  
Vietnam  
Spain  
Turkmenistan  
Trinidad and  
Tobago  
Madagascar  
Slovenia  
Kenya  
Slovakia  
Syrian Arab  
Republic  
Ethiopia  
Côte d'Ivoire  
Cameroon  
Benin  
Serbia  
Portugal  
Netherlands  
Montenegro  
Lithuania  
Korea  
Peru  
Morocco  
Mexico  
Mauritius  
Libya  
Zimbabwe  
Togo  
Kyrgyzstan  
Iraq  
Tanzania,  
South Sudan  
Rwanda  
Nigeria  
Israel  
Ireland  
Iran  
Hungary  
Greece  
Guatemala  
Ecuador  
Dominican  
Republic  
Niger  
Germany  
Namibia  
Mozambique  
Mongolia  
Mauritania  
Mali  
Georgia  
France  
Colombia  
Bangladesh  
Azerbaijan  
Argentina  
Venezuela  
Ukraine  
Finland  
Denmark  
Cyprus  
Croatia  
Malawi  
Liberia  
Belgium  
Austria  
Tunisia  
Sri Lanka  
Paraguay  
Panama  
Nicaragua  
Nepal  
Lesotho  
Guinea-Bissau  
Gambia  
Australia  
Armenia  
Uruguay  
Latvia  
Gabon  
Congo  
Chad  
Canada  
328  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Central African  
Myanmar  
Jamaica  
India  
Republic  
Burundi  
Burkina Faso  
Botswana  
Angola  
Honduras  
Guyana  
Cuba  
Costa Rica  
Chile  
Brazil  
Bolivia  
Algeria  
Compiled by the author  
Figure 3 - Maps of variables with the boundaries of clusters with the grouping attribute inExpl  
329  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Most of all, we will be interested in the first picture, which reflects the distribution of  
countries according to the level of exploitation. We clearly see that the C4 cluster was  
formed by countries where this problem is especially strong. Since the representatives of the  
cluster are heterogeneous in other parameters, the averaged statistics of average indicators  
does not describe it well enough. We also see that the second place in terms of exploitation  
is occupied by the Middle East cluster C5. In addition, in the cluster-interlayer C1, the  
contour of a subcluster with high exploitation rates is outlined. The main “marker” of a high  
utilization rate is the volume of outgoing money transfers.  
Clusters completely free from the problem have not been formed; The lowest rate of  
exploitation, contrary to many theoretical models linking the level of trafficking with the  
level of development of civil and human rights institutions, is in the countries of the stratum  
cluster, where these institutions are not well developed. In those countries where the indices  
of equality and freedoms are high, also contrary to expert estimates, the level of exploitation  
is at its maximum. Thus, today developed civil institutions are unable to reduce the rates of  
trafficking. The general low level of social inequality also does not contribute to a decrease  
in exploitation; on the contrary, it is under a more equitable social system that exploitation  
is carried out more often.  
Another important fact that we can observe on the maps: the lack of a constant  
relationship between trafficking and migration. If the correlation for the countries of the  
Middle East can be traced, then it is absent in the cluster of leaders in terms of exploitation:  
the number of migrants here is negligible. If we relate the flows of incoming migration to the  
level of traffic, then one would expect that the European cluster would become the leader;  
however, this does not happen and the hypothesis of a direct relationship between these  
phenomena must be rejected.  
Another important conclusion that we can draw from distribution maps is the  
definition of the main "pivot points" of clustering, that is, the basic principles of difference.  
This can be done by examining the boundaries of the clusters. The variables that define them  
will be the most important. Such variables for inExpl include: the number of old people and  
children, life expectancy, literacy, rural population, general population, gender inequality.  
330  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
If we graphically display the clusters on the map, marking the highest risk of  
exploitation in red and the lowest in green, we get a map of the distribution of countries  
depending on the structure of the risk of exploitation (Figure 4).  
Figure 4 - Geographic distribution of countries by level of exploitation  
Further, the indicator of outgoing traffic is analyzed in a similar way. Our task is to  
identify key donor countries and their properties. We divide all countries from the sample  
into 5 clusters (Figure 5).  
Figure 5 - Map of clusters with grouping attribute sourceCountry  
331  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Average indicators for clusters are presented in Table 3.  
Table 3 - Table of average values of cluster variables with grouping attribute sourceCountry  
migrant nat  
Gender  
Population  
0-14  
Cluster Frequency Origin  
Inflation stock  
income Corruption Inequality  
C 1  
25,69%  
27,08%  
24,31%  
17,36%  
5,56%  
1,757  
1,923  
1,543  
3,6  
3,32  
3,22  
3,12 5404  
2,5 974  
38,51 0,4037  
27,1 0,5908  
27,85  
42,52  
16,32  
23,32  
22,19  
C 2  
C 3  
C 4  
C 5  
1,14 11,27 24701  
5,23 4,41 5925  
-13,08 54,65 21408  
64,34  
0,1121  
34,36  
0,3193  
1
47,25 0,2836  
Sec  
Fertility Life  
rate  
Adolescent  
Social  
expectancy Happiness support Literacy fertility  
education Population  
25+  
65+  
C 1  
2,4  
72,45  
61,26  
79,84  
72,5  
5,404 1,264  
4,071 0,856  
6,318 1,399  
5,383 1,325  
6,128 1,228  
Rural  
96,6  
71,3  
99,8  
98,5  
99,2  
54,3  
97,3  
10,8  
34  
59,5  
31,1  
89,9  
78,7  
65,1  
6,78  
3,05  
17,28  
9,16  
C 2  
C 3  
C 4  
C 5  
4,735  
1,621  
2,135  
2,277  
76,41  
14,5  
2,97  
Pop  
Remit  
growth Undernourish Tuberculosis pop  
Mobile Unemployment inflow  
Death  
6,61  
C 1  
1,2  
2,496  
0,281  
0,826  
3,564  
8,98  
23,86  
2,85  
6,36  
5,04  
124,1  
196,3 62,76  
15 27,42  
123,8 47,22  
12,7 10,44 149,2  
35,3  
114,5  
71,4  
9,38  
6,88  
10,56  
5,19  
2812  
1657  
3966  
11193  
1678  
C 2  
C 3  
C 4  
C 5  
8,76  
9,66  
8,61  
2,83  
123,3  
121,4  
4,26  
Compiled by the author  
Cluster 1 (C1) - 26% of observations. This group has mainly average economic and  
demographic indicators, while there is a fairly high gender inequality, morbidity,  
unemployment, literacy, the number of migrants here is relatively small.  
Cluster 2 (C2) - 27% of observations. Average inflation and unemployment rates, low  
NPI are combined with very high rates of morbidity, corruption, threat of hunger, and  
inequality indices. It has the highest birth rate, relatively high mortality, high population  
growth, and short life expectancy. In all respects, this cluster resembles the “disadvantaged”  
countries from the previous stage of the analysis.  
Cluster 3 (C3) - 24% of observations. These countries are characterized by low  
inflation, high NPI, low fertility and population growth, a large number of migrants. It has  
the highest death and unemployment rates, the highest life expectancy, the highest level of  
happiness and the lowest level of corruption.  
332  
REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 34, 2021  
Yulia A. Siluyanova // Trafficking in persons: cluster analysis of sending and receiving countries, 317-340  
Cluster 4 (C4) - 17% of observations. In many respects, it has average indicators,  
stands out with a record high level of social support, the highest inflation, mortality is quite  
high here, there are quite a lot of elderly people, and low unemployment rates. This is where  
the largest flow of remittances from migrants comes from abroad.  
Cluster 5 (C5) - 6% of observations. There is negative inflation, a huge number of  
migrants, high NPI and favorable social indicators.  
Here again we see a division in accordance with the current geopolitical situation.  
Two “safe” clusters have emerged, one “dysfunctional” and a layer between them.  
Further attention is paid to how the socio-economic statistics of different countries  
are distributed among groups. For this, the distribution graphs for each variable are  
considered (Figure 6).  
The conventional wisdom is that we expect to see the highest traffic rate in the  
dysfunctional” C2, but this is not happening. This cluster ranks second in terms of outbound  
traffic, while the leader is the “average” C4. It is this cluster that makes up the core of donor  
countries. We also observe a group with the lowest possible level of the indicator, that is, a  
cluster that is not actually engaged in the “export” of live goods. These are C5 - Middle  
Eastern oil economies, which are second in terms of exploitation. Cluster C3, that is,  
European countries, has a very low rate of outbound traffic.  
In order to make sure that our assumptions about the composition of the clusters are  
correct, consider the results of the grouping (Table 4).  
This time the grouping turned out to be less geographically accurate: the first cluster  
contains the countries of Latin America, Africa, and some Asian states. The second cluster is  
still almost entirely composed of African countries. The third and fifth clusters with  
industrialized developed economies as well as the oil powers of the Middle East have  
changed little. But cluster 4, which brought together the most problematic states, expanded.  
It is noteworthy that the volume of remittances, this time incoming, again became a  
marker of high traffic.  
Geographically distributing the clusters on the map, we obtain a scheme for the  
distribution of the risk of outgoing trafficking, where the most active donor countries are  
shown in red, and the countries with the lowest level are indicated in green (Figure 7).  
333