Crime Crisis and Economic Growth
Crime Crisis and Economic Growth
The Keep
Spring 2022
Recommended Citation
Thapa, Ankita, "Crime, Crisis and Economic Growth: An Investigation of Socio-Economic Determinants of
Crimes in the Indian States" (2022). Masters Theses. 4921.
https://thekeep.eiu.edu/theses/4921
This Dissertation/Thesis is brought to you for free and open access by the Student Theses & Publications at The
Keep. It has been accepted for inclusion in Masters Theses by an authorized administrator of The Keep. For more
information, please contact [email protected].
CRIME, CRISIS AND ECONOMIC GROWTH: AN
By
Ankita Thapa
Department of Economics
December 2021
This thesis is submitted in partial fulfillment of the requirement for the award of a master’s degree
Thesis Committee
1
ABSTRACT
This paper investigates the impact of socio-economic conditions on five major crime heads
from 2001-2019 using a panel data set for the Indian states. The paper focus on the great recession
of 2008-09, economic growth of the states, and deterrence variables. The paper employed two
estimation procedures: panel Fixed-Effect and two-stage least square-fixed effect (2SLS-FE). The
2SLS-FE is preferred over the fixed effect method, where poverty is treated as an endogenous
variable with higher education and social sector expenditure as instrumental variables. A dummy
variable is used for the period of the great recession. A square of state GDP per capita is used to
find evidence of a non-linear relationship between crime and economic growth. Three deterrence
variables, namely conviction, arrest, and police strength are used in the model.
The study found that the great recession positively impacted total crime, violent crime, and
crime against women. The economic growth (State’s GDPs) positively affects total crime,
economic offenses, and crime against women. The study also found evidence of an inverted U-
shaped curve (non-linear relationship) between three crime categories and economic growth. The
deterrence variables do not seem to have a deterrence effect on crime in the Indian states. However,
their coefficients are highly significant but positive. Poverty and unemployment only seem to
2
ACKNOWLEDGEMENT
I would like to thank my committee supervisor Dr. Ahmed Abou-Zaid for his supervision,
guidance, and continuous support that led to the successful completion of this thesis. His advice
and suggestions have immensely inspired me to pursue my goal and finally achieve it. I would also
like to thank the committee members of this thesis Dr. Ali R. Moshtagh and Dr. Linda Ghent. I
sincerely appreciate all the professors for giving me learning opportunities in the past three years.
I would specially like to mention Dr. Mukti Upadhyay for believing and guiding me during
my first year in Eastern Illinois. I couldn’t have done it without his tremendous support.
I would also like to thank my parents and sisters, who always sent me love and prayers from
overseas and encouraged me to do better. Finally, I am thankful to my loving husband Sahil for
providing comfort and support during difficult times and encouraging me to keep going on this
3
TABLE OF CONTENTS
ABSTRACT………………………………………………………………………………………2
ACKNOWLEDGEMENT……………………………………………………………………….3
LIST OF TABLES……………………………………………………………………………….6
LIST OF FIGURES………………………………………………………………………………6
CHAPTER ONE……………………………………………………………………………...7-10
1. Introduction…………………………………………………………………………………......7
CHAPTER TWO……………………………………………………………………………11-19
2. Crime in India………………………………………………………………………………….11
CHAPTER THREE…………………………………………………………………………20-26
3. Literature Review……………………………………………………………………………...20
CHAPTER FOUR…………………………………………………………………………...27-34
4
4. Empirical Strategy……………………………………………………………………………..27
CHAPTER FIVE……………………………………………………………………………35-43
CHAPTER SIX……………………………………………………………………………...45-48
6. Conclusion ……………….……………………………………………………………………45
APPENDIX……………………………………………………………………………………...49
REFERENCES………………………………………………………………………………….51
5
LIST OF TABLES
LIST OF FIGURES
Figure 4: Total crime Vs. Police strength in selected States (2001 and 2008)……………………17
6
CHAPTER ONE
1. INTRODUCTION
There is not a single country in the world that is not affected by the evils of crime. Still, some
countries face more detrimental consequences than others. Although the level of development is
beginning to lay less and less influence on curbing crime in today's world, there is still a consensus
that developed countries encounter few crime incidents (especially amongst the developing
countries). Nevertheless, one cannot deny that easy access to weapons, illegal drugs, and advanced
technology has made it easier for criminals to commit crimes nowadays. Crimes are as rampant in
a developed country as in a developing economy. However, the former has more resources to
India is a developing country of 1.3 Billion people with soaring crime rates. In past years
various crimes, especially crimes against women, skyrocketed. Despite having one of the highest
GDP growth rates, India lags in deterring criminal activities, which can be traced back to a lack of
In 2019, India registered 3,225,701 for IPC crimes and 1,930,471 SLL crimes. Crimes
increased by 1.6 percent from the previous year, with 32.6 percent of the total IPC crime accounted
for crime against body. Crime against women increased by 7.3 percent, and property crime
increased by 6.5 percent in 2018 (National Crime Records Bureau, n.d.). The annual GDP growth
of India in 2019 was 4.04, the unemployment rate was 5.27 percent, and the poverty rate was 6.7
The crime and socio-economic statistics differ inside the states of India. Goa has the highest
GDP per capita ($7,032), and Bihar has the lowest ($2,395) (Knoema.Com, n.d.). In terms of total
7
crimes, Uttar Pradesh (12.2% of total national crime), Maharashtra, and Kerala are the top three
states. Assam recorded the highest rate of violent crime, 86.4 per 100,000. Madhya Pradesh
recorded the highest rape cases, 16.3 percent of total rape cases in India (Times now, n.d.).
Nagaland has the least crime rate recorded at 51.8 incidents per 100,000 compared to the national
average of 241.2. It is also one of the safest states for women, with only 0.8 incidents of rape per
However, it is difficult to ascertain the impact on crime during the recession. During the great
recession, some countries experienced high crimes while in some it declined. The great recession
of 2008-09 impacted India's trade by dropping exports to 16 percent. The Indian government then
introduced three stimulus packages between December 2008 to February 2009, equivalent to US$
245 Billion (3.5 % of GDP). The GDP dropped to 6.7 percent in 2008, then recovered soon to 8.5
percent in 2009-10 ("2008 Global Financial Crisis," Iyer). However, criminal activities did show
an uptick. The IPC crime rose by 5.2 percent, crime against body increased by 4.8 percent, crime
against property increased by 8.8 percent, and economic crime rose by 3.6 percent compared to
the previous year. Crime against women, which includes molestation (4.3%), cruelty by husband
India would seem uniform to the outside world, but the enormous cultural, social and
geographical disparity makes it unique. Therefore, the choice of incorporating a panel data analysis
is appropriate, as proved by the statistics mentioned before. The socio-economic scenario and
crime rates differ amongst the states, and hence the union and states’ government work accordingly
1
See appendix for total IPC crime map of India
8
1.2 Research Questions
The paper seeks to analyze the impact of socio-economic conditions and factors on significant
crime categories: total crime, violent crime, property crime, economic offenses, and crime against
women. It also seeks to determine the deterrence effect of deterrence variables that is arrest,
1. Does the Great Recession of 2008-09 influence the five given crime categories for
Indian states?
2. Does the states' economic growth have any impact on all crime categories? Do they
The motivation behind this study arises from the lack of research on the effect of economic
growth, great recession, deterrence on crime rates, with a particular focus on Indian states. Most
of the past research focused on the effect of crime on economic growth, not the other way around.
The idea of exploring the non-linear relationship between economic growth and crime does stem
from a research paper written by Nayebyazdi (2017). He incorporated the Kuznets curve propelled
by Simon Kuznets in his paper. However, there is a deficiency of papers analyzing the non-linear
relationship between economic growth and crime. Therefore, this paper is the first such attempt
about India.
Moreover, several previous studies are at the national level. This study will explore the crime
and other factors at the state level and can deliver much detailed information.
9
Although the great recession period is significantly studied, the main focus was on financial
markets, monetary policies, and international trade. The literature has a dearth of studies on
recession and its impact on crime, specifically in the Indian subcontinent. This study will equip
the existing literature with empiric-based estimations on the relationship between recession and
crime.
The rest of the study is divided into six chapters. Chapter two elaborates on the crime trends
in India and Indian states. Chapter three presents a literature review of previous theoretical and
empirical studies and research. Chapter four describes the empirical strategy followed in the study,
along with some preliminary tests and their results. Chapter Five will elaborate on the summary
statistics and discuss the results from the two estimation techniques applied in the study. Finally,
Chapter six concludes the study with summary, limitations, and future recommendations.
10
CHAPTER TWO
2. CRIME IN INDIA
The Indian Criminal procedure Code (CrPC) provides a complete punishment procedure
under the penal laws. It enforces and administers the Indian Penal Code (IPC) and other criminal
laws.
In the Indian Criminal Procedure Code (CrPC), there are two categories under which different
crimes are recorded. One is a cognizable offense, and another is a non-cognizable offense. The
police have to take prompt action when a complaint is filed under a cognizable crime. They have
to visit the crime scene, gather evidence and information, apprehend the criminal and produce the
offender before the court of law. All of this takes place without a court-mandated warrant. It
Under a non-cognizable offense, the courts summons the offender and, if needed, involve
police investigation with the permission granted by the magistrate. It often includes less severe
crimes. Under the cognizable offense, there are two sub-categories of law: the Indian Penal code
(IPC) and the Special and Local Law (SLL) under which crimes are registered. The IPC is the
official criminal code (a document) intended to cover all aspects of criminal law and breach of
law. It consists of 23 chapters, 511 sections and covers a wide range of offenses with prescribed
penalties and the punishments for the respective crimes (Code of Criminal Procedure, 1973, 1974).
Under SLL, crimes that violate unique acts or laws framed by parliament, and state or local
government are registered. It covers specific acts and regulations like indecent representation of
women, transplantation of human organs, lottery prohibition, sale of liquor, etc. The SLL crime
11
data are not included in this study as the legal procedure followed under such crimes is different
Criminal
Procedure Code
(CrPC)
Cognizable Non-cognizable
Offence Offence
IPC
SLL
The National Crime Records Bureau (NCRB) has maintained the crime database annually
since 1986. It is an open database created to provide law enforcement agencies and the general
public with information. The data is available at the national, state, district, and metropolitan levels
In the executive summary by NCRB, there are six broad classifications of crimes under the
2. Crimes against property: It includes dacoity, its preparation & assembly, robbery
12
5. Crimes against women: It includes rape, dowry death, cruelty by husband and
procuration of minor girls, selling and buying of girls for prostitution, abetment of
summary, n.d.).
In 2019, the NCRB recorded a rise in overall crime by 3 percent. Crime against women rose
by 7.3 percent, economic offenses by 6.1 percent and, property crime by 6.5 percent in the same
year. However, a rise in criminals' arrest and conviction rates by police and courts are also observed
As of now, India has 28 states and eight union territories. In October 2019, Jammu and Kashmir
was reorganized into two union territories: Jammu & Kashmir and Ladakh. The difference between
a state and a union territory is that the former has its own government while the Union government
of India governs the latter as a federal territory. But this paper includes Jammu and Kashmir as a
state because the crime data before 2019 recognize it as a state. Only Delhi is added to the study
because it experiences high crime rates like any other state and serves as the national capital of the
country.
Similarly, Telangana, formed in 2014, is not included in the study as the data is insufficient.
I did not add any other union territories of India because the data are not chronologically
maintained, and the crime rates are negligible compared to all the states.
13
Figure 2: Total crime growth in the Indian States
Figure 2 provides the total crime growth rates in Indian states in 2019 from 2018. Uttar Pradesh
recorded the highest total crime growth rate of 12.8 percent. Similarly, many states like Madhya
Pradesh, Maharashtra, and Rajasthan experienced a rise in total crime by more than 12 percent in
2019. We can also ascertain that crime is higher in bigger states from the map. North-Eastern states
have less crime growth rate than all other Indian states, but the crime rate rose from the previous
year in these states, too, except Sikkim. For example, in 2018, Nagaland recorded 6.9 percent
growth in total crime in 2018 and 7.1 percent in 2019. In Delhi, actual crime rises by 12.6 percent
in 2019, approximately equivalent to crime rates in the big states mentioned above.
14
2.3 Crime and Recession
Figure 3 provides the comparative analysis of five states and one union territory from different
zones for 2001, 2008, and 2009, representing the period before and after the Great Recession.
DELHI GUJARAT
2001 2008 2009
11.7
11.7
11.5
2001 2008 2009
10.2
10.2
10.0
10.9
10.8
10.8
10.1
10.0
9.1
9.0
8.9
8.8
8.8
8.7
8.2
7.8
7.8
8.5
8.4
8.3
8.3
8.3
7.9
7.8
7.7
7.5
4.6
KERALA MAHARASHTRA
2001 2008 2009 2001 2008 2009
12.2
12.2
12.1
11.7
11.6
11.6
11.2
11.2
11.0
10.0
10.0
9.8
9.7
9.6
9.4
9.2
9.4
9.3
9.3
9.2
9.3
9.3
9.2
9.0
9.0
8.7
8.6
8.3
8.2
8.1
15
MIZORAM UTTAR PRADESH
2001 2008 2009 2001 2008 2009
12.1
12.1
12.0
10.6
10.5
10.5
10.4
10.2
10.2
10.1
10.1
7.7
7.6
7.6
7.2
9.9
7.0
7.0
9.5
9.5
9.1
5.2
5.1
5.1
5.1
5.0
4.8
4.7
4.7
4.6
The different graphs of the states given here help to visualize the growth in various crime heads
in pre-and post-crisis years. In Delhi, the growth in total crime rate was consistent in pre-and post-
crisis years. Nevertheless, when we look at the crime heads separately, all categories showed a
spike in crime growth rates. Property crime fell in 2008 compared to 2001, but it rose sharply in
2009. There was not much difference in the total crime rate of these states, but crime against
Uttar Pradesh and Maharashtra saw a rise in economic crimes during the Great Recession. The
rise in economic offenses post-recession was expected; however, it declined in Kerala, Mizoram,
and Gujarat. Mizoram experienced less growth in crime as compared to the other states because it
16
2.4 Crime and Police strength
Figure 4: Total Crime vs. Police Strength in selected states (2001 and 2008)
206243
178129
171233
168996
152211
123808
123050
118076
110620
107840
103847
103419
54384
52834
49350
48747
46017
45703
36742
35666
3535
2686
2246
1989
Data Source: NCRB, BPRD; Graph: Author's construction
Figure 4 demonstrates the total number of crimes reported and police strength of the respective
states in 2001 and 2008. Uttar Pradesh and Maharashtra are two states that experience the highest
criminal activities despite the highest number of the police force. Delhi had a high total crime rate,
but the police force was almost equivalent to the reported crimes in both years. Kerala had
comparatively less police strength against total crime. Only Gujarat and Uttar Pradesh had less
It was noted that Mizoram had registered less total crime in 2008 than in 2001. On the other
hand, Mizoram had more than 3,535 police officers compared to the 1,989 total crime cases
registered in 2008. Mizoram had more police officers than the total crime in 2001 as well.
17
2.5 Crime and Economic Growth
Delhi Gujarat
13.00
12.00
12.50 11.90
12.00 11.80
SGDP
SGDP
11.50 11.70
11.00 11.60
10.50 11.50
11.00 11.50 12.00 12.50 13.00 10.00 10.50 11.00 11.50 12.00 12.50
TOTAL CRIME TOTAL CRIME
Kerala Maharashtra
12.60 12.80
12.70
12.40
12.60
12.20 12.50
12.40
SGDP
SGDP
12.00
12.30
11.80 12.20
12.10
11.60
12.00
11.40 11.90
10.00 10.50 11.00 11.50 12.00 12.50 10.50 11.00 11.50 12.00 12.50
TOTAL CRIME
TOTAL CRIME
SGDP
12.20
7.50 12.00
11.80
7.40
11.60
7.30 11.40
7.20 11.20
0.00 5.00 10.00 15.00 9.50 10.00 10.50 11.00
TOTAL CRIME TOTAL CRIME
18
Figure 5 demonstrates the relationship between the state GDP per capita and total crime of
selected states from 2001 to 2019. The primary purpose of the graph is to look for evidence of a
non-linear relationship between the two variables, which is based on the idea of the Kuznets curve
put forward by Simon Kuznets in 1955. He states that as income rises, inequality in economic
The graphs depict that in Delhi and four states, as SGDP rises, crime rates also tend to rise;
however, several minor turning points in the curves are detected. Although the curves eventually
19
CHAPTER THREE
3. LITERATURE REVIEW
The following chapter is divided into two sections elaborating past literature on crime and
economic factors.
a) Unemployment
During a recession, the level of unemployment in a country starts to rise, exacerbating poverty
and inequality that might lead people to resort to illegal activities for their livelihood.
Gary S. Becker gave one of the earliest theoretical works about economic conditions and crime in
one of his seminal works in 1968. Later, his theory, known as the economic theory of crime, states
that the opportunity cost of engaging in criminal activities reduces when unemployment increases.
Some people will be willing to risk conviction and find criminal activities more attractive to
generate more immediate financial/material gain. The rise in unemployment will eventually lead
In a study concerning the US, Cantor and Land (1985) explained two effects for economic
downturns (implied by unemployment) and crime. In the first effect, the positive relationship
between the unemployment rate and crime is due to the motivation effect that unemployment
motivates people to commit crimes. The negative relationship is due to the opportunity effect in
the second effect. The crime rate may decrease due to increased unemployment as the opportunity
to commit the crime is reduced because of increased guardianship by the unemployed population.
The relationship between unemployment and crime suffers from inconsistencies in empirics
throughout the literature. Researchers found positive, negative, and no relationship between the
20
two variables. Moreover, the studies on crime are primarily based on the United States and
In New Zealand (Papps & Winkelmann, 2000), Sweden (Blomquist & Westerlund, 2014), and
Canada (Janko & Popli, 2015), there was no significant relationship between unemployment and
crime.
In a study of crime in the United States by Phillips and Land (Phillips & Land, 2012),
unemployment increased vehicle theft at the city level. Their state-level analysis found that
unemployment positively impacts theft, robbery, and vehicle theft. However, the relationship
between theft and crime was weak at the state level. In EU-28 countries, Ayhan and Bursa (Ayhan
& Bursa, 2019) found a positive relationship between unemployment and crime. In EU-28, the
unemployment nor inequality significantly influences crime; however, the level of development
and the GNP per capita was strongly correlated with property crime.
In a developing country, Tabar and Noghani ( 2019) studied Iran from 1997-2013 and found a
significant positive relationship between unemployment and crime. A rise in unemployment will
increase the crime rate by 0.44 units in Iran. In a panel of 16 developing countries from different
regions, Anser et al. (2020) employed system GMM to estimate the relationship between various
macroeconomic factors and crime rates. They found that unemployment increased crime rates by
In India, the relationship between crime and unemployment was mixed as well. Cui and Hazra
(2017) used Johansen cointegration and granger causality to confirm the presence of a relationship
between unemployment and crime rate. Their study involved time-series data from 1991-2015
21
with GDP per capita, unemployment, and inflation as dependent variables. Bharadwaj ( 2014)
found that a 10 percent increase in unemployment would reduce property crime by 0.4 percent.
His research concluded that poverty is a more valid factor influencing crime, especially property
crime.
b) Poverty
Poverty alone cannot influence crime, and not all people living in poverty would commit a
crime. There are multiple channels and circumstances through which poverty can impact or
aggravate crime. Rising poverty increases the inequality between rich and poor, reducing the
opportunity for better education, increasing unemployment, and worsening living conditions. All
Bourguignon (1999) defined crime as a social cost of poverty and inequality where the cost
arises from the loss of material, expenditure on crime prevention, and punishment. He estimated
that in the US and Latin America, the cost would be up to 3.7 percent of GDP and 7.5 percent,
respectively. Bhardwaj (2014) found that a 10 percent rise in poverty would significantly increase
However, a panel of 16 countries Anser et al., (2020b) did not find a significant relationship
between poverty headcount ratio and crime rates in system GMM estimates. However, in the
Variance dynamic response (VDA) study, poverty will exert the most significant share to influence
crime, and in the IRF study, poverty increases the crime rate from 2015 to 2024. Another study in
Mexico (Martinez, n.d., 2012) found that a rise in poverty increased property crime by 2.37 units
in 2000 and 4.36 units in 2005. Also, violent crime rose by 5.32 units in 2000 and 3.99 units in
2005 in Mexico.
22
c) Economic growth
Researchers have mixed views on economic growth and crime, and most of the studies are
based on crime as one of the factors influencing the country's economic growth and not the other
way round. However, the studies have two varying theories regarding the impact of economic
growth on crime rates. One is that a bad economy might force a person to commit crimes to make
ends meet, and another one is that a thriving economy may increase crime rates. A wealthy
neighborhood might experience more property crime and robberies because of the availability of
In an empirical study by Oliver (2002), no significant relationship was found between the crime
rate and economic growth in the USA between 1960 to 1998. However, during 2007-2010, the
economist ( 2011) found that crime rates dropped as the United States dealt with the economic
downturn.
Mulok et al. (2016) used the ARDL approach to test the relationship between crime and
economic growth in Malaysia for 1980-2013. They found that crime rates increased by 0.93
percent when GDP per capita increased by one percent. They also found a two-way causal
relationship between the two variables. Similarly, Dutta and Husain (2009) found that growth in
the economy led to an increase in crime rates in India from 1999 to 2005. An increase in per capita
state product increases the crime rate by 28.69 units in their SURE model. They argued that this
could be due to the rising inequality resulting from GDP growth. Another study found the Kuznets
curve in relation to economic growth and crime. Nayebyazdi (2017) found that in the European
Union, during the early stages of economic growth, inequality increases and leads to more crime,
but it starts to decline over time. His findings laid the basis for analyzing a non-linear relationship
23
d) The Great Recession
The past literature on crime and recession found that it would increase crime rates. However,
the crime rate fell during 2008-2009, especially in the US and developed nations. It was speculated
that an improved criminal justice system, better social safety nets, and cultural changes might have
restricted the impact of the recession on crime rates (Rosenfeld, 2014). Furthermore, the crime
itself is related to many factors that have counterinfluence, making some types of crime less likely
to occur (Uggen, 2012). However, the recession would impact crime rates through rising
unemployment in the long term. It was found that entering the labor market during a recession
increases the arrest rate by 10 percent and conviction rate by 4 percent than entering during a more
conducted by United Nations Office on Drugs and Crime (UNODC) for selected 15 countries
during the economic downturn of 2008-2009 found that seven countries experienced a rise in at
least one crime type. Robbery and motor vehicle theft more often occurred during this period,
consistent with criminological theories that violent and property crime increases during economic
stress. Homicides (violent crime) are positively associated with economic downturns in the case
of countries with a high level of violence like Brazil, El Salvador, and Jamaica (UNODC, n.d.).
The lack of empirics regarding recession and India further emphasized the significance of this
study. Viswanathan (2010) found that crisis spread in India through three channels: financial,
trade, and confidence. Through the confidence channel that is the sentiments of consumers and
investors, there was a decline in consumption by households and businesses, further reflected in
24
Before the crisis, India's GDP growth was recorded at plus 9 percent for three consecutive
years, but in 2008-09 it dropped to 6.7 percent. However, India recovered quickly in 2010 and
recorded a growth rate of 7.7 percent, attributed to private investments, considerable savings, a
young population, high productivity growth, and flexible fiscal and monetary policies (Dasgupta
Furthermore, Indian banks and financial institutions did not include mortgage back securities
and credit default swaps. The service exports were not hit hard as merchandise exports, remittance
was recorded at US$ 46.4 billion in 2008-09, and FDI was still high US$27.3 Billion (Aiyar, 2009).
There is much literature on recession, its causes, impact, and policy measures in India; however,
research that can provide evidence of a relationship between the great recession and crime in India
In simple terms, deterrence means fear of punishment. A person is restrained from committing
a crime when the thought of consequences like arrest, trial, conviction, and sentence comes to his
mind. The foundation of the deterrence theory is based on the work of 18th-century philosophers
Jeremy Bentham(2017, [1789]) and Cesare Beccaria (1963 [1764]), which states that punishment
will deter a criminal to commit crime further and the knowledge of the severity of punishment will
further deter a potential offender. They perceived a criminal as a rational individual who weighs
More spending on security measures like the increasing police force, patrolling, enhancing the
police force with the latest technology might deter a potential criminal and help reduce crime rates.
25
Moreover, the severity of punishment, quick arrest, and conviction of criminals prevent recidivism
A report produced by Congressional research services (CRS) (Finklea, n.d.) indicated that there
are all possible results of the impact of the police force on crime- it can increase, decrease, and
have no effect at all. An additional officer prevents 2.9 violent crimes and 16.23 property crimes
In a panel study of England and Wales from 1989 to 1996, a one percent rise in police force
leads to a 1.32 percent fall in vehicle crime and a 0.38 percent fall in property crime (Witt et al.,
1999).
In India from 1999 to 2005, the presence of more police officers resulted in 0.02 units decline
in crime rates. However, an increase in conviction and arrest rate led to a positive and significant
rise in crime which the author explained as a consequence of corruption and malpractice in the
Indian jail system that alleviates criminal tendencies (Dutta & Husain, 2009). Nevertheless,
Amaral et al. (2014) found mixed results on the presence of police force on different crimes at the
district level from 1990-2007. They found that property and economic crimes are associated with
an increased police force. However, violence and crime against women decreased. Also, arrest
rates declined the crime rates in all categories, that is, 0.19 percent property crime, 0.32 percent
violent crime, 0.06 percent economic crime, and 0.21 percent crime against women.
researchers add deterrence variables like arrest rate, punishments/incarceration rate, conviction
rate, police force to their model. I followed the literature mentioned above and chose three
26
CHAPTER FOUR
4. Empirical Strategy
The study employs detailed state panel data that have been constructed on the incidence of
various types of crime, socio-economic variables, and deterrence variables for the 2001-2019
period. The study aims to examine the impact of the great recession on various types of crime and
find evidence of any non-linear relationship between crime and economic growth for Indian states.
A total of 29 entities (states)2 are included in the study. Only one union territory is included in the
study, New Delhi, as it functions as a state and reports high crime rates compared to other union
territories.
The data is collected from various government data portals. Crime and police data are collected
from NCRB (National Crime Records Bureau, n.d.) and BPRD (Data on Police Organizations:
Bureau of Police Research and Development, Government of India, n.d.), data for socio-economic
variables are collected from OGD (Open Government Data Platform India, n.d.), (Government of
India, Ministry of Education, n.d.) and Reserve bank of India (DBIE-RBI : DATABASE OF
Four sets of variables are used in the study and transformed into logarithmic forms due to the
2
See appendix for list of states included in the study
27
a) Crime Variables: Five types of crime variables are included in the analysis as dependent
variables. The total crime (Totalcrime) is the incidence of all types of IPC crime reported within
a state in a year. Violent crime (VCrime) includes all incidence of murder, attempt to murder,
rape, kidnapping and abduction, dacoity, riots, arson, and dowry deaths. Property crime (PCrime)
includes the incidence of theft, robbery, and burglary. Economic offense (EconOff) includes
incidence of cheating, criminal breach of trust, counterfeiting, and crime against women
(CrimeAwomen), includes incidence of rape, dowry death, immoral traffic, assault, or insult to
b) Deterrence variables: Three types of deterrence variables are used for the analysis. First, the
total number of persons arrested (TArrest) within a state in a year for different crime heads are
collected for the research. Secondly, conviction (conv) is included as the number of cases disposed
of by police and court for all crime heads in a year. Lastly, Police strength (polstr) is added to the
model, including active civil police and district armed force strength. The active civil police have
direct contact with the general public, and the district armed police force have their own police
officers with higher firearm competence. The state armed force is not included in the study as it
does not have any contact with the public and only works during public events like civil unrest,
c) Socio-economic Variables: I used unemployment, poverty, post-crisis, and state GDP per capita
the total unemployed per thousand. Poverty (pov) is measured as the percentage of people below
the poverty line. Post-crisis is a dummy variable for the great recession of 2008-09, and State GDP
28
per capita (SGDP) is the proxy for the state's economic growth. The post-crisis (postcrisis) dummy
variable will take 0 for 2008 and 2009 or 1 for 2001-2007. As the variables constitute different
units, all are transformed into their logarithmic form for a more straightforward interpretation.
d) Control Variables: There are other variables that might influence crime in the Indian states.
Education and expenditure on social services might negatively affect crime rates (Benoît &
Osborne, 1995; Mittal, n.d. ; Hjalmarsson & Lochner, n.d.). Therefore, my models have included
education and social sector expenditure as control variables. I used enrollment in higher education
(HEdu) as a proxy for education. It includes the number of pupils in general education courses
like bachelors and master’s in arts, Science and Commerce, and professional education courses
Social sector education (SoSecExp) is the expenditure by the government in billions (Rs.) in
social services like education, sports, arts and culture, family welfare, water supply, sanitation,
housing, social security, and welfare, nutrition, and expenditure during natural calamities.
I employed two estimation procedures in this analysis. First is the standard fixed effect
estimation, and the second is the Two Stage Least Square-Fixed Effect (2SLS-FE).
Besides the results from the Hausman test, there is another reason why the fixed effect is
preferred over the random effect in both models. The fixed effect model is appropriate while
analyzing the impact of variables that vary over time. However, in the presence of time-invariant
characteristics, the fixed effect will remove its impact from the model (Torres-Reyna, n.d.). Some
time-invariant variables that might influence crimes in the Indian states like gender, culture,
29
religion, and languages can be measured by applying the Random effect. However, these variables
are not well maintained and hardly available in India. Since it is difficult to measure them and their
exclusion might lead to omitted variable bias. Applying the Fixed effect method in such case will
control all the omitted time-invariant differences and remove the bias.
The 2SLS is an extension of the OLS method, applied when the error term is correlated with
the independent variables. This problem arises due to an endogenous variable (pov) on the left-
hand side of the equation which is correlated to the error term. In such case, instrument variables
are used in two stages. First, the predicted value of problematic variable is obtained by
implementing instrument variables. In the second stage, the predicted value of the endogenous
variable is used to estimate the desired model. Since the data is panel, 2SLS-FE estimation
procedure is applied.
The endogeneity problem stems from three reasons - omitted variables, simultaneity, and
measurement errors. As mentioned before, there are omitted variables in the models which might
impact crimes. Moreover, some independent variables are highly correlated to each other, which
can generate biased results. The endogeneity problem can be proved by the correlation matrix
Furthermore, the results from Panel fixed effects (Results section) indicated unexpected signs
for poverty and deterrence variables which indicated model specification problems.
a ) Correlation Test: One of the linear model assumptions is that the regressors are exogenous; that
is, they are uncorrelated with the error term. However, if any regressors correlate to the error term,
30
Table1: Correlation Matrix
Variables (1) (2) (3) (4) (5) (6) (7) (8) (9)
(1) totalcrime 1.000
(2) Unemp -0.239 1.000
(3) pov 0.022 0.007 1.000
(4) HEdu 0.728 -0.223 -0.838 1.000
(5) conv 0.927 -0.212 -0.023 0.656 1.000
(6) PolStr 0.829 -0.168 0.020 0.612 0.388 1.000
(7) TArrest 0.952 -0.242 0.001 0.594 0.723 0.702 1.000
(8) SoSecExp 0.859 -0.190 -0.885 0.132 0.780 0.638 0.131 1.000
(9) SGDP -0.269 -0.053 -0.332 -0.208 -0.211 -0.302 -0.248 -0.061 1.000
The results from the correlation matrix suggest that crime is strongly correlated with all the
variables except poverty, unemployment, and SGDP. However, there is multicollinearity between
independent variables. Poverty is strongly correlated to education and social sector expenditure
and employing them in the analysis would give inconsistent and biased results. The best way to
tackle the problem of multicollinearity is to drop the problem variables; however, HEdu and
SoSecExp are also strongly correlated with the dependent variable and dropping them might give
b) Sargan-Basmann Test: To choose appropriate instrument variables (IVs) for poverty, two
-It should be correlated with the endogenous variable. Cov (z, pov) 0
-It should be uncorrelated to the error term () and exogenous. Cov(z, )=0
After the fulfillment of these requirements, an identification test must be done. Here, z
represents the instrument variable. Since the equation has more than one IV, it is overidentified.
The equation can be just identified (one IV for one endogenous variable), overidentified (more
than one IVs) but never to be under-identified. The overidentification restriction test computes
31
Sargan's (1958) and Basmann's (1960) Chi2 statistics for regression estimates through instrumental
The p-value fails to reject the null hypothesis. Therefore, the instruments are valid.
c) Durbin-Wu-Hausmann Test: After selecting the appropriate IVs, 2SLS is executed to perform
the endogeneity test. To confirm the endogeneity in the model, a Durbin-Wu-Hausman test (1954)
(1974) (1978) is employed, which is performed after the execution of 2SLS regression with HEdu
The test's null hypothesis suggests that the variables are exogenous (pov).
The null hypothesis is rejected at 5 percent significance. It means that the variable poverty is
endogenous.
32
d) Hausman Specification Test: The Hausman specification test is generally used to select fixed
and random-effects models. The Hausman test (1978) is applied to select the appropriate model
for both estimation techniques. It examines if the individual effects are uncorrelated with other
regressors in the model. If the individual effects are correlated with any regressor, we cannot use
random effect as it is no longer BLUE. In such cases, the fixed-effect model is more efficient and
reliable. The null hypothesis of the Hausman test states that errors are not correlated with the
regressor, and the alternative hypothesis states that they are correlated. A random effect is used if
we fail to reject the null hypothesis, and a fixed effect is used to reject the null hypothesis.
(Equation 2)
(Equation 4)
All major crime heads have the exact model specification for all estimation techniques.
33
a) Fixed Effect Model:
b) 2SLS-FE :
Where X represents all the exogenous variables included in the equation [2]. The instrument
variables are HEdu (Higher Education) and SoSecExp (Social Sector Expenditure).
However, the Stata program runs the two-stage equation as a single equation, and the two steps
In the second stage equation, pov (poverty)'s predicted value is used instead of the observed
poverty rates.
34
CHAPTER FIVE
The results section is divided into three sub-sections. In the first sub-section, summary statistics
of all variables are given. In the second sub-section, the Fixed effect panel estimation results are
The descriptive statistics of variables are given in table 5 below. The maximum number of
crimes is 353,131 recorded by Uttar Pradesh in 2019, and the lowest is 443 recorded by Sikkim in
2003. The considerable deviation of 80323.65 indicates that the data is more spread out to its mean
value. Similarly, most crime variables have a massive deviation from the mean, which is expected.
Some Indian states experience relatively more minor crimes than others.
Violent crime was maximum in Uttar Pradesh in 2018, recorded at 65,155, whereas the
minimum of 51 was recorded in Sikkim in 2003. Delhi experienced the maximum number of
property crimes in 2019, recorded at 251,548 and only 100 in Sikkim in the same year. The
property crime averages around 17,602. The state of Tripura recorded 0 economic offenses in the
year 2012. However, Sikkim, which has low crime rates with respect to major crime heads,
surprisingly recorded the maximum number of economic offenses, which is 27,071. Uttar Pradesh
again tops the chart in crime against women. The state recorded 59,853 cases in 2019. Sikkim
again recorded only 24 cases related to crime against women in 2003. However, it should be noted
that many crimes, especially crimes against women, are underreported as most women fear
retaliation from the assailant, some experience shame, and some do not want to damage family
35
honor. Also, many petty crimes like theft and robbery of less valuable articles are not reported.
The long legal process and bureaucratic red tape-ism discourage people from reporting crimes.
The maximum number of students enrolled in higher education is 6.5 million in Uttar Pradesh.
The lowest enrollment recorded was 3,850 in Sikkim in 2003, which is unsurprising as there is a
massive difference in the area and population of both states. The maximum unemployment rates
36
per thousand is 21.35 recorded in Nagaland in 2019, and the minimum is 0.55 recorded in Gujarat
in 2018. The average unemployment rate for all states is around 4.14 per thousand.
The highest poverty rate was recorded in Odisha in 2009 where 57.2 percent of the population
was living below the poverty line. Jammu and Kashmir had the lowest poverty rate in 2003 at
3.48% below the poverty line. Apart from Odisha, many other states like Bihar, Jharkhand, and
The maximum number of criminals convicted is recorded at 352,340 in Uttar Pradesh in 2019,
and the minimum conviction was recorded at 23 in Arunachal Pradesh in 2017. However, we can
see that the arrest rate is relatively high compared to the conviction rate. Almost 3.3 million people
were arrested in Maharashtra in 2017, and a minimum of 246 people were arrested in Sikkim in
2014.
The civil and armed police strength has an average of 41,000 for all states. Maharashtra police
have the maximum number of police force, which is recorded at 209,397 in 2017. Sikkim police
The average state GDP per capita is around 78,621.88 rupees, equivalent to $1046.72 with a
standard deviation of $834. The maximum state GDP per capita was recorded in Goa in 2015 at
4,39,756 rupees ($5854.62), whereas the minimum was recorded at 9,888.25 rupees ($131.61) in
Bihar in 2002.
In the case of social sector expenditure, Uttar Pradesh spent the maximum amount in 2019.
The state government spent approximately 1,577 Billion rupees which is equivalent to US$ 2
Billion. The Sikkim government spent the minimum amount in 2001, which is 2.6 billion rupees,
37
5.2 Fixed Effect Panel Estimation
The result from Panel fixed effect estimation is given in table 6. It is given to present a reference
point for acknowledging differences in the two estimation techniques used in this paper. The
unexpected results are shown in FE estimation also laid the basis for undertaking an additional
38
The poverty variable has an unexpected negative sign across all major crime heads except
property crime. However, only three are significant. Violent crime, economic crime, and crime
against women reduce as poverty increases by one percent. On the other hand, unemployment has
Higher education reduces total crime, violent crime, and economic crime significantly by 0.02
percent, 0.06 percent, and 0.13 percent, respectively. Both conviction and total arrest increase
crime significantly in all crime categories except one in each. However, a percent increase in
conviction reduces violent crime by 0.14 percent. When we look at Total arrest, all crime
categories are positively related to the arrest rate. A percent increase in arrest rate would increase
total crime by 0.32 percent, violent crime by 0.52 percent, property crime by 0.57 percent, and
crime against women by 0.26 percent. Only the coefficients of police strength have expected
Similarly, expenditure in social sector would significantly increase total crime, violent crime,
and crime against women by 0.17 percent, 0.19 percent, and 0.22 percent, respectively.
The dummy variable for great recession (post-crisis) is insignificant in the analysis. However, all
A percent increase in SGDP reduces violent crime by 3.04 percent but increases property crime
and economic crime significantly by 4.8 percent and 3.2 percent, respectively. The negative sign
of the coefficient of SGDP and the positive sign of SGDP 2 indicates a U-shaped curve for violent
crime. It indicates that as the economy grows further, violent crimes will eventually rise. On the
other hand, property and economic crimes would reduce as the economy grows as both exhibits
SGDP does not significantly affect total crime and crime against women in this model.
39
5.3 2SLS-FE Estimation
The results from the 2SLS-FE estimation are given in table 7. Due to unusual results from the
The poverty variable in the model is significant at 5 percent for a total crime only. However,
as seen in the previous estimation, no crime heads turn up with an unexpected sign. The poverty
variable is endogenous in this model, with higher education and social sector expenditure as the
instrument variables. A percent rise in poverty would increase total crime by 0.31 percent. In the
40
(UNDP) in 2019, India is selected as one of the ten countries that reduced their MPI values. India
has lifted 271 million people out of poverty between 2006 and 2016. Amongst the Indian states,
Jharkhand has reduced the incidence of poverty from 74.9 percent in 2005-06 to 46.5 percent in
2015-16 ((MPI) | Human Development Reports, n.d.). This reduction in Multidimensional poverty
might placate crimes in the Indian States. However, 28 percent of the total population still lives in
poverty in India in 2019 and have potential to influence crimes. Another explanation for a
significant positive total crime could be that many poverty programs are not directed towards the
target groups. For example, instead of assisting poverty-stricken people who are more prone to
commit crimes, many economic programs mainly focus on women, children, physically
challenged, and elderly. Youth, unemployed, uneducated, and unskilled are almost always left out
of these programs and hence might resort to criminal activities to fulfill their needs.
Unemployment was unable to affect crime significantly in the previous model, but it
significantly affects total crime by reducing it by 0.08 percent. All other crime variables are still
insignificant.
The results were expected for unemployment as, throughout the literature, the effect of
unemployment is mixed. There has been positive (Ayhan & Bursa, 2019) (Mir Mohamad Tabar &
Noghani, 2019), negative (Bharadwaj, 2014) (Cantor & Land, 1985), or no relationship (Janko &
Popli, 2015) (Krohn, 1976) found between unemployment and crime by various researchers.
unemployment. Unemployed are available to provide better security and protection to their
properties and families than people at work, creating less opportunity for criminals to participate
in any criminal activities. Hence, it will create a negative effect on crime rates.
41
The three deterrence variables show similar signs and significance in both models. The
conviction rate significantly increases total crime by 0.23 percent, economic crime by 0.46 percent,
and crime against women by 0.27 percent. Only violent and property crimes exhibit expected signs
but are insignificant in the analysis. Conviction does not always mean jail time for many criminals
in India. The convicted can appeal at a higher court and be out on bail if it is a bailable offense
(section 389 CrPc, n.d.) allowing the offender to commit crimes further.
The rise in crime against women can be explained by the presence of relationship between the
offender and the victim. Most of the culprits are husbands, family members or relatives. Women
are forced to retract complaints after the conviction of the assailant many times, leading to a cycle
of harassment for women. Moreover, corruption and bureaucracy within the Indian policing
system and the legal system can keep the culprit out of the system for a long time.
The arrest rate is also positively significant in affecting most major crime heads except
economic crime. A percent increase in arrest would increase total crime by 0.29 percent, violent
crime by 0.45 percent, property crime by 0.67 percent, and crime against women by 0.20 percent.
The results align with the findings of Dutta & Husain, n.d. (2009), who found a positive
relationship between conviction and crime, and arrest and crime for India.
An arrest does not necessarily mean conviction and incarceration of that person. The inefficient
and weak criminal justice system makes trials longer, affecting appropriate and sound judgment.
A person can be acquitted due to lack of evidence, be out on bail, or be released after a warning
for petty crimes, which can act as an anti-deterrent for criminals. Moreover, corruption, especially
nepotism, makes it worse. Criminals with high profile families or relatives like politicians, people
in business, or powerful positions have leeway to circumvent the legal system according to their
suiting. Also, in many Indian states and their respective policing system, corruption runs from the
42
lowest level to the highest position, making disposal of criminal cases very rigid. The source of
this corruption is the power vested in police officers (Lamani & Venumadhava, 2013). Criminals
can walk freely by simply paying a bribe to the officials, which further creates a haven for criminals
The strength of the police force is insignificant in affecting all crime heads, but the signs are
as expected and in line with previous findings (Mello, 2019) (Witt et al., 1999). An increase in the
SGDP and SGDP2 were insignificant in the previous model for the total crime. Here, both
significantly impact total crime, economic offense and crime against women. A percent rise in
SGDP would increase total crime by 3.8 percent, and a percent rise in SGDP2 would decrease total
crime by 0.18 percent. Similarly, economic offense would rise by 6.1 percent as SGDP increase
and starts to decline by 0.3 percent as SGDP 2 increases further. Crime against women rise by 5.2
percent and tend to decrease by 0.23 percent as SGDP continue to increase. This indicates the
presence of an inverted u-shaped curve for total crime, economic offense, and crime against
women against economic growth. As the economy grows, crime rates will rise and decline after
reaching a certain point. A possible explanation for this relationship could be the increasing income
inequality. The rise in income does not applied to all the sections of society here. States like
Gujarat, Maharashtra, Haryana, Kerala, and Punjab are still the wealthiest states while, Bihar,
Assam, Odisha, Rajasthan, Uttar Pradesh are at the bottom. There is a presence of inequality within
these states too. This social injustice might encourage people to commit violent crimes and
property crimes against the affluent members of society. In lack of a potential victims, criminals
43
always target the weakest members of the community: children or women to vent out their
As we know, economic growth can increase the inequality gap, which may impact crime
positively. However, over years as the economy grows and inequality reduces, crime starts to
decline (Nayebyazdi, 2017). In this estimation, all three significant crime categories exhibit an
inverted U-shaped curve, which is different from the previous one where the three crime categories
The dummy variable post-crisis is positive and significantly impact crime in this estimation
technique. During the great recession (2008, 2009), total crime rose by 0.09 percent, violent crime
by 0.10 percent, and a crime against women by 0.11 percent compared to the pre-crisis period,
other things being constant. It is in line with the findings of UNODC regarding developing
countries. They found that violent crime will increase in countries more prone to violence. During
recessions, unemployment is high. For a densely populated country like India, the unemployment
rate of more than 5 percent in past two decades is a considerable number (Statista, n.d.). The crime
will likely go up in such situation due to frustration amongst unemployed. However, it contradicts
current literature that expects a negative relation between the great recession of 2008-09 and crime
rates due to increased guardianship. The studies were based on advanced economies only.
Therefore, it is not appropriate to expect the same outcome for developing countries.
44
CHAPTER SIX
6. Conclusion
This paper analyzed the impact of socio-economic conditions, specifically poverty, great
recession, deterrence, and economic growth, on five major crime categories for the Indian states.
The sample covers 28 states and one union territory of India, ranging from 2001 to 2019. A total
of 10 independent variables are used with five dependent variables. I applied two different
estimation techniques: Panel Fixed effect and 2SLS-FE, which produced different outcomes. The
fixed effect panel estimation did not provide enough information and suffered from an econometric
issue. Poverty turned up with unexpected significant negative signs, and unemployment was
insignificant in the first estimation. Therefore, 2SLS-FE is used to overcome the endogeneity issue
1. The poverty variable is positively related to the crime categories and significantly impacts
total crime.
2. Unemployment reduces total crime through the opportunity effect first put forward by
3. The great recession of 2008-09 had a significant impact on crime. It increases total crime,
4. There is a non-linear relationship between the economic growth of states and crime. Total
crime, economic crime, and crime against women have an inverted u-shaped curve
45
indicating a rise in crime at an initial increase in SGDP and then drop as the economy
grows.
5. The two deterrence variables, conviction and police strength, have a significant impact on
crime. Conviction and rise in police force increase the crime rate in Indian states due to the
widespread corruption in the policing system. However, the deterrence effect does not take
place.
As we know that India suffers from poverty, and at least two-thirds of the population live in
poverty (Poverty in India, n.d.). For a vast population, that number is enormous. We cannot always
conclude that only poor people commit crimes, but poverty likely creates a premise for criminal
behavior. Apart from property crime, poverty can also influence a person to commit violent crimes
There are numerous policies and programs in effect that aim to eradicate poverty. However,
the reach of those programs is uncertain and does not always reach the target population. The
incompetent governance, especially in rural areas and widespread corruption, obstruct such
programs' desired results. So, education is the key here which can enhance human capital.
Investment in human capital accumulation would target the problem areas like unemployment and
poverty and reduce crimes in future. Furthermore, to overcome the ignorance of general
population, government officials should educate people about poverty programs through various
The Indian legal system should be more versatile. Incorporating new laws and procedures
according to the present crime scenario can enhance the effectiveness of the justice system. The
major drawbacks of Indian legal system are the lack of transparency, lack of coordination and
46
corruption. This should be addressed to have consistency in the working of the legal system. More
fast track courts, judges and lawyers should be appointed for speedy trials and conviction.
The great recession's impact was not as severe as in the developed countries, but the rising
inequality does pose a significant threat. Again, investment in education, raising the minimum
wage, and generating more employment opportunities will help shorten the gap.
Secondly, to curb crime, the monetary penalty should be high and corrupted officials should be
monitored regularly. For ex-offenders, programs and schemes should be formulated to find
employment, education, and accommodation that further enables them to live dignified lives and
There are certain limitations to this research. First is the availability of the data. The census of
India is taken every ten years, and the latest data available was from 2011. Some essential variables
that can influence crime, like inequality and population, are not included because the data are
unavailable. Secondly, Union territories and a newly formed state, Telangana, are not included as
data points were not maintained for the years selected in this research. Their inclusion might affect
For future research, high crime states and low crime states can be studied separately to have a
more concrete idea about which policies or program works better in each case. If data permits,
including data points from districts, towns, and villages can also give a new approach to this
research. Many crime studies suggested including a lag variable of crime as the past crime rates
47
As the deterrence variables to do not pose deterrent effect on crime variables in the Indian states
in this paper, including a lag of police strength variable might present different results. The number
of police force from last year can influence crime in the present year. However, restricted
knowledge in advanced econometrics could not facilitate the idea in this paper. The same model
can be replicated by including these additional variables which can further enhance the findings of
this paper.
48
APPENDIX
LIST OF STATES
49
Figure 6: Total IPC Crime map of India 2008
50
REFERENCES
2008 global financial crisis: What government overdid or didn’t do. (2021, April 15). The Indian
Express. https://indianexpress.com/article/explained/global-financial-crisis-2007-08-
indian-economy-lehman-brothers-fall-5359333/
https://core.ac.uk/download/pdf/296912781.pdf
Aiyar, S. S. A. (2009, September 13). India weathers 12 months of financial crisis. The Economic
Times. https://economictimes.indiatimes.com/swaminathan-s-a-aiyar/india-weathers-12-
months-of-financial-crisis/articleshow/5005007.cms?from=mdr
Amaral, S., Bandyopadhyay, S., Bhattacharya, S., & Sensarma, R. (2014). Crime and social
conflict in India. The Economics of Peace and Security Journal, 9(1), Article 1.
https://www.epsjournal.org.uk/index.php/EPSJ/article/view/191
Ayhan, F., & Bursa, N. (2019). AVRUPA BİRLİĞİ ÜLKELERİNDE İSTİHDAM VE SUÇ
https://doi.org/10.35408/comuybd.574808
650–659. https://doi.org/10.1080/01621459.1960.10483365
Benoît, J.-P., & Osborne, M. J. (1995). Crime, Punishment, and Social Expenditure. Journal of
51
Bharadwaj, A. (2014). Is poverty the mother of crime? Empirical evidence of the impact of
Cantor, D., & Land, K. C. (1985). Unemployment and Crime Rates in the Post-World War II
https://ncrb.gov.in/sites/default/files/CII%202019%20SNAPSHOTS%20STATES.pdf
Crime and the Great Recession. (n.d.). Retrieved December 1, 2021, from
https://inequality.stanford.edu/sites/default/files/Crime_fact_sheet.pdf
Dasgupta, D., & Gupta, A. S. (2010). India: Rapid Recovery and Stronger Growth after the Crisis.
In The Great Recession and Developing Countries (pp. 235–301). The World Bank.
https://doi.org/10.1596/9780821385135_CH06
https://bprd.nic.in/content/62_1_DataonPoliceOrganizations.aspx
DBIE-RBI : DATABASE OF INDIAN ECONOMY. (n.d.). Retrieved September 29, 2021, from
https://dbie.rbi.org.in/DBIE/dbie.rbi?site=home
Do recessions increase crime? (n.d.). World Economic Forum. Retrieved December 1, 2021, from
https://www.weforum.org/agenda/2015/03/do-recessions-increase-crime/
52
Dutta, M., & Husain, Z. (n.d.). DETERMINANTS OF CRIME RATES CRIME, DETERRENCE
Dutta, M., & Husain, Z. (2009, January). Determinants of crime rates: Crime Deterrence and
Explained: Bail After Conviction Section 389 CrPc. (n.d.). Retrieved December 1, 2021, from
https://legalserviceindia.com/legal/article-5785-explained-bail-after-conviction-section-
389-crpc.html
https://www.unodc.org/documents/data-and-
analysis/statistics/crime/GIVAS_Final_Report.pdf
https://doi.org/10.2307/1913827
Hjalmarsson, R., & Lochner, L. (n.d.). The Impact of Education on Crime: International Evidence.
Research Report, 8.
India GDP per capita by state, 2020—Knoema.com. (n.d.). Knoema. Retrieved December 9, 2021,
from https://knoema.com//atlas/India/ranks/GDP-per-capita
Kuznets, S. (1955). Economic Growth and Income Inequality. The American Economic Review,
45(1), 1–28.
53
Mello, S. (2019). More COPS, less crime. Journal of Public Economics, 172, 174–200.
https://doi.org/10.1016/j.jpubeco.2018.12.003
Mir Mohamad Tabar, S. A., & Noghani, M. (2019). Unemployment and crime in developing
countries: A meta-analysis in Iran. Crime, Law and Social Change, 72(3), 327–338.
https://doi.org/10.1007/s10611-019-09823-y
Mittal, P. (n.d.). Social Sector Expenditure and Human Development of Indian States. 21.
National Crime Records Bureau. (n.d.). Retrieved April 20, 2021, from https://ncrb.gov.in/en
November 27, K. D. N. D., December 7, 2020 ISSUE DATE:, November 27, 2020UPDATED:,
& Ist, 2020 13:27. (n.d.). Nagaland: A Safe Haven. India Today. Retrieved December 9,
safe-haven-1744595-2020-11-27
Open Government Data (OGD) Platform India. (n.d.). [Text]. Open Government Data (OGD)
Poverty in India: Facts and Figures on the Daily Struggle for Survival. (n.d.). Retrieved December
Rosenfeld, R. (2014). Crime and the Great Recession: Introduction to the Special Issue. Journal
Statistics New | Government of India, Ministry of Education. (n.d.). Retrieved September 29, 2021,
from https://www.education.gov.in/en/statistics-new?shs_term_node_tid_depth=384
Viswanathan, K. G. (2010). The Global Financial Crisis and its Impact on India. 9, 23.
54
With 16% of entire nation’s rape cases, Madhya Pradesh records highest number of rapes: NCRB
now/crime/article/with-countrys-16-of-rape-cases-madhya-pradesh-records-highest-
number-of-rapes-ncrb-report/538037
Witt, R., Clarke, A., & Fielding, N. (1999). Crime and economic activity. A panel data approach.
World Bank Open Data | Data. (n.d.). Retrieved July 23, 2021, from https://data.worldbank.org/
Wu, D.-M. (1974). Alternative Tests of Independence between Stochastic Regressors and
https://www.princeton.edu/~otorres/Panel101.pdf
The 2019 Global Multidimensional Poverty Index (MPI) | Human Development Reports. (n.d.).
Unemployment rate in India 2007-2018. (n.d.). Statista. Retrieved January 4, 2022, from
https://www.statista.com/statistics/271330/unemployment-rate-in-india/
55