PRELIMINARY AND INCOMPLETE
DO NOT QUOTE
Political Representation and Crime:
Evidence from India’s Panchayati Raj*
October 2009
Lakshmi Iyer
Harvard Business School
Anandi Mani
University of Warwick
Prachi Mishra
Research Department, IMF
Petia Topalova
Asia-Pacific Department, IMF
Abstract
We examine the impact of political representation for women on crimes against them, using data from a
decentralization initiative in India. We exploit variation in the timing of the implementation of this
initiative across states to identify the impact. We find that having female leaders in local government
induces a large and significant increase in reported crimes against women. There is no significant effect
on crimes against men or gender-neutral crimes. The effects are not significant for crimes where underreporting is not a serious concern. Overall, our results show that having female leaders at the local level
lends a strong voice to women, resulting in a sharp increase in the reporting of crimes against them.
JEL Classification Numbers: O10, O17, O20
Keywords: Crime; decentralization
*
Authors’ E-mail addresses: liyer@hbs.edu; a.mani@warwick.ac.uk ; pmishra@imf.org; ptopalova@imf.org
The views expressed in this paper are those of the author(s) and do not necessarily represent those of the IMF or
IMF policy. We are grateful to our colleagues at the IMF for helpful comments. Maya Shivakumar, Filipa Jorge and
Heisnam Thoihen Singh provided excellent research assistance.
1. Introduction
One of the most fundamental aspects of civil society is the assurance of personal safety and fair treatment
under the law for all citizens. Indeed, fair representation of the interests of all citizens, minority or
majority, is one of the professed aims of all democracies. However, protecting the rights of disadvantaged
groups has remained a challenge in many countries, especially developing ones. Such politically
disadvantaged groups can include population minorities or those who are politically under-represented,
such as women in most developing countries. Individuals from such groups are often the target of a whole
gamut of injustices ranging from everyday indignities such as verbal abuse to serious crimes including
murder and even genocide. In this paper, we examine the effects of a specific policy instrument,
mandated political representation for political minorities, on crime outcomes against them. We exploit the
changes in local political representation of women and other disadvantaged groups induced by a major
decentralization initiative in India to examine this issue.
Why would leaders from a minority community matter for its welfare? It could simply be that their life
experience as a minority group member gives them a different perspective, which could affect both policy
making and implementation. As the most recent U.S. Supreme Court appointee Sonia Sotomayor has
observed “I further accept that our experiences as women or people of color affect our decisions…Hence, one must accept the proposition that a difference there will be by the presence of women or people of color on the bench. Personal experiences affect the facts that judges choose to see.”1 Furthermore, unlike
appointed judges, political leaders can only retain office only if they keep voters happy. This provides
minority leaders more incentives to be responsive to the needs of their largest constituency, i.e. minority
voters. In this paper we use the term “minorities” to refer to political minorities or groups which are
under-represented in the political sphere.
The effects of such mandated political representation have been previously studied in the context of
targeted public expenditures (Pande, 2003) and local public good outcomes (Chattopadhyay and Duflo,
2004). However, it is important to study crime outcomes, over and above specific public good outcomes,
for two reasons. First, any policy intended to give voice to minorities can only be regarded as a success if
it enhances their acceptance and status in society as a whole. Crime against minorities is a broader
measure of their welfare because it serves as a useful barometer of their socio-economic status and level
of empowerment in broader society, beyond the reach of government. Second, the experience of having
political minority leaders itself can change the attitudes towards the minority community.2 In fact, better
access of the minority community to jobs or education, through political representation or reservations,
may also affect the majority’s attitude towards them, which would be reflected in the crime rates. These
attitudinal changes may often be more negative when public good outcomes for minorities see significant
improvement.
Our paper is the first to examine the effects of a mandated political representation for minorities on
crimes against them. The related economics literature on crime has focused on socio-economic factors
influencing it, ranging from local inequality and economic shocks to peer effects and family structure.
Some papers have analyzed the impact of specific policy interventions, such as increased police hiring
(Levitt, 2007; Di Tella and Schargrodsky, 2004), prison overcrowding (Levitt, 1995), and, more
1
2
Judge Mario G. Olmos Memorial Lecture in 2001, University of California, Berkeley (Sotomayor, 2002).
See Beaman et al (2008) on this in the context of attitudes to women leaders in India. In the U.S., there is
anecdotal evidence to suggest that the election of Barack Obama as the first African-American President in office is
having an effect on race relations in the country.
controversially, legalized abortion (Donohue and Levitt, 2001). Further, most of these existing analyses
have focused on overall crime levels, or on broad crime categories such as property crimes or violent
crimes. Very few are concerned with crimes against specific sections of society3 – and none have
examined the role of political power for minorities on crime outcomes against them.
To study the effect of political representation for minorities on crime outcomes against them, we outline a
basic framework that considers the incentives of three key sets of agents in our context: criminals, victims
and law enforcement officials 4. We identify two distinct channels for how politician identity may matter
for crime outcomes. The first is that the presence of a minority group politician may increase the
likelihood of action by law enforcement officials in cases of crimes against that group’s members. This
would materialize because of the effects of the minority leader’s perspectives and priorities in policy
making and implementation, as outlined earlier. The increased likelihood of punishment would serve as a
deterrent on the actions of a potential criminal against a minority group member. Second, minority
victims may feel more emboldened to report crimes against them, when they perceive that the politician
in power is more sympathetic to their concerns.5
In examining the overall impact of these two channels on the incentives of criminals and victims, we see
two effects that go in opposite directions. On the one hand, the greater possibility of punitive action can
deter crimes against minorities. On the other, both the possibility of punitive action and the presence of a
sympathetic political leader in office can encourage minorities to report crimes against them more often.
It is therefore an empirical issue to see which of these two effects dominate in actual crime outcomes. A
negative effect of minority political representation on crimes suggests a dominance of the deterrence
effect over the reporting effect, while a positive effect of minority political representation on crime
suggests dominance in the opposite direction.
One obvious empirical hurdle in studying the effects of political reservation for minorities is the
endogeneity in which segments of the electorate choose to elect a minority group politician. Typically,
this would make it hard to disentangle the effects of having a minority political representative in office on
outcomes from the effects of unobservable characteristics of the electorate. However, we are able to
address this issue using data from a unique experiment with Panchayati Raj (PR) – or local government
system in India. Beginning in 1993, an amendment to the Indian constitution mandated greater
decentralization and empowerment of district and village level local governments. The same amendment
also provided for political reservations for marginalized groups (women and people of Scheduled Castes
and Tribes(SC/ST).) A certain number of seats in the district and village councils were to be set aside for
these groups.
While there are other cases of countries implementing political representation for
minorities, the Indian case is an important one to study given that it is regarded as among the most
radical, with regard to minority political representation. (Grofman and Lijphart (1986)). 6
We start by conducting a state-level analysis, by exploiting the fact that although the legislation was
introduced uniformly across all states, the actual implementation date varied considerably across states
3
The exceptions include Miguel (2005) on the effect of economic shocks on witch-killing and Donohue and Levitt
(2001a) on the link between the racial profile of police offers and the racial pattern of arrests.
4
For an analysis of the implicit contractual conditions of political competition under which representation for
minorities can influence their policy outcomes, see Pande (2003).
5
For instance, Chattopadhyay and Duflo (2004) find that women in Indian villages are more likely to attend local
government meetings and speak up during them if the local government leader is a woman. Hoff and Pandey (2007)
offer experimental evidence of how low caste identity may affect minority expectations and behavior: They find that
lower caste children perform worse in games when their caste identity is made public by (upper caste) experimenters
– specifically because they anticipate discrimination against themselves, and not due to a loss in self-confidence.
6
Please refer to section 3 for details on the relevant provisions of the Indian Panchayati Raj system.
(for various reasons detailed below). Thus, different states implemented these reforms at different times,
and had different pre-existing levels of crimes against vulnerable groups. Our identification strategy thus
compares areas where political positions were reserved for women (say) to areas where such positions
were not reserved, controlling for other observable characteristics.7
Using data on the timing of introduction of political representation for women for our state level analysis,
we find that it has large positive and significant effects on overall crimes against women, as well as on
specific types of crime such as kidnapping of women and rape. These results are robust to the inclusion
of state and year fixed effects, a host of controls for economic, demographic and political variables, the
strength of the police force and even state-specific linear trends. Equally interesting is the fact that we
find no significant effects whatsoever on any categories of crime not specifically targeted against women
– such as kidnapping of men, crimes against property or crimes against public order. This strongly
suggests that there is no overall deterioration in law and order conditions or policy changes other than the
political representation that are driving our results.
Given the opposing effects of politician identity on crime outcomes outlined above, we interpret this
increase in observed crimes against women as driven by a net increase in reporting of crime rather than in
actual crime incidence. This is consistent with the evidence for categories of crime such as murder,
where the reporting bias is least likely to be a problem. For this category, we find a negative effect of
politician representation on crime occurrence. In addition, we find that there is no increase in crime
outcomes of a less serious nature, where reporting is less likely to be influenced by politician identity.8
Finally, we look at the effect of political representation at higher levels of government, which is typically
much lower than the thirty three percent mandated for local government positions, and was not changed
by this reform. We find that having a woman as the head of the state (Chief Minister) – as opposed to a
local government representative -- has a negative and significant effect on crimes against women.
Women at intermediate levels of political office have a positive (but insignificant) effect on crime against
women, as with the local government female representatives. In the context of our framework outlined
earlier, these findings are consistent with a scenario where leaders’ power for punitive action is stronger
at higher levels of government, even as their impact on crime reporting is invariant across levels. Hence,
a (more powerful) woman Chief Minister may have the greatest deterrence effect on crime, whereas local
level women leaders may simply enable women to voice their protest against crimes. This is particularly
plausible under the Indian Panchayati Raj initiative where authority for law and order was not devolved to
the local government.
In ongoing work, we’re exploring this and other hypotheses regarding crimes against women using
district level data. We are also examining the effects of political reservations for Scheduled castes and
Scheduled Tribes for crimes against those minorities.
The paper is laid out as follows. Section 2 lays out a simple framework to capture the various channels
through which political representation can affect observed crime outcomes and arrests. Section 3
provides details on the institutional details of the Panchayati Raj (PR – local government system) in India
and Section 4 describes the data and empirical strategy. Section 5 presents our results and robustness
checks so far, and offers a cohesive interpretation of our findings. Section 6 describes additional ongoing
work and concludes.
7
The second source of variation, which we plan to exploit in ongoing work, is that even among the states which
implemented this reform, different areas were reserved for this different groups.
8
Given the category definitions, crimes registered as IPC (Indian Penal Code) crimes are regarded as more serious
than those registered as SLL(Special and Local Laws) crimes.
2. Political Decentralization and Crime
2.1 Crime: Occurrence, Reporting and Actions Taken
We outline a framework similar to Donohue and Levitt (2001a) that considers the incentives of three
parties relevant to a crime situation: criminals, victims and the policing authority. First, a potential
criminal decides whether to commit a crime or not. Next, when a crime occurs, the victim must decide
whether to report it. Finally, for a crime that is reported, the police must decide how much time and effort
to devote to investigating the crime and bringing the criminal to justice. The schematic diagram below
outlines the sequence of events.
Crime Opportunity
No Crime Occurs
Crime Occurs
Crime Not Reported Crime Reported
Action taken No Action taken
Let us consider the factors that affect the decisions of each of these parties, in sequence. First, the
probability of a criminal act (C) would be a decreasing function of both the probability that the victim
reports it (R), and probability of punitive action being taken against him (A). 9
C = C (R, A)
This brings us to the victim, and the factors that affect his/her decision to report a crime.
R=R(A; S; c (IP))
We identify three key factors: the probability of action being taken (A), the Seriousness (S) of the crime
and the victim’s cost of reporting (c). With regard to A, if there is little hope of justice being delivered,
crime victims will see no point in reporting crime. The parameter S captures the extent of psychological
We include R and A as separate arguments because criminals may care about reporting independent of its eventual
effect on punitive action against them too, given its potential adverse effects on their reputation within the
community.
9
or emotional utility of the victim (or informant) from letting others know about injustices meted out to
them – and this is likely to be higher for more serious crimes.
The parameter c captures the costs of reporting a crime to the police, which can be economic and
psychological. Economic costs may simply be the cost of getting to the nearest police station.
Psychological costs may include trauma from recounting the event, fear of retaliation by the criminal,
frustration with the legal authority, or shame – all of which may be particularly relevant for minorities.
For instance, women may face distress and shame in cases of sexual harassment, rape etc. or a person of
SC/ST background may fear retaliation if he reported crimes against him by upper caste folks. Both
groups may face frustration with an unsympathetic law enforcement authority.
We conjecture that a victim’s cost of reporting a crime also depends on the identity of the politician in
power IP. If the identity (caste/gender) is the same as that of the victim (IV), then the victim is more likely
to report the crime, since he or she feels that they are more likely to be listened to. 10 A similar
relationship can hold if the police officer is more likely to record the crime (rather than simply dismiss the
complaint or harass the victim) if IP = IV .
Finally, we posit that the probability of action being taken against criminals by law enforcement officials
(A) is an increasing function of the seriousness of the reported crimes (S), as well as other factors such as
citizens’ awareness and the resources and commitment that the state has to law enforcement.
A = A(IP ; S, Other factors)
Most importantly, we conjecture that the probability of punitive action is higher when there is a match
between the caste/gender identity of the politician IP and that of the crime victims.
Given these assumptions about the effects of politician identity IP on action against criminals A and the
cost of reporting c, a match between the identity of the politician and victim (IP =IV) would have an
impact on both C and R. On the one hand, a higher A will induce a negative effect on crime occurrence,
C. On the other hand, it will also induce a positive effect on crime reporting, R. Furthermore, IP would
also increase R by reducing minorities’ psychological cost of reporting a crime, c. The overall effect on
observed (or reported) crime outcomes is therefore ambiguous.
One point to note however, is that for the most serious categories of crimes against minorities
(where the reporting bias is likely to be weakest), the negative crime reduction effect must
dominate when IP =IV. For other crimes where there is no such match between politician and
victim identity, IP ≠IV, political representation for minorities should generate a positive/zero
effect on observed crime outcomes. Thus, we generate the following testable implications.
2.2 Testable Implications
10
This is supported by evidence from Chattopadhyay and Duflo(2004), which shows that women are more likely to
attend village meetings and speak up during them if the village leader is a woman. Economic costs of reporting
crimes may also be higher for minorities if they have to make repeated visits to do so, when faced with
unsympathetic police personnel.
(1) For crimes against minority victims (women and/or members of the Scheduled Castes or
Scheduled Tribes), the overall effect of having a minority group politician in power is ambiguous.
A positive effect on such crimes indicates a net increase in crime reporting while a negative effect
indicates a net reduction in actual crimes.
(2) For the most serious categories of crimes against minorities (where the reporting probability is
highest), the negative crime reduction effect is likely to dominate. For the least serious crimes
against minorities, there should be little effect of politician identity on crime outcomes.
(3) Minority political representation should have no effect on categories of crime where victims do
not belong to the minority group.
It is an empirical question whether the relevant political representative with the power to influence action
being taken against criminals is at the village, district, the state/national level – or one or more of these.
On the one hand, a Chief Minister (head of the state government) has greater influence over the police,
both because of their prominence, and the fact that the state government is the one who controls the
police. On the other, the local politician is more likely to know the local police personnel and be
accessible to crime victims. We prefer to let the data speak for itself on this matter, by conducting our
analysis for women representatives at both the local and the state level.
3. Crime, Policing and Political Representation in India
3.1 Crime Procedures
The Code of Criminal Procedure provides the basis for the criminal justice system in India. This code
specifies that all information given to the police must be reduced to a written report by the police officer,
read and signed by the informant. After this “First Information Report” or FIR has been filed, the police is
required to investigate the crime, and maintain detailed police diaries of the progress of the investigation.
During such investigation, the police may question or arrest any suspects.11
At the conclusion of the investigation, the police must deliver to the magistrate a police report which
details the results of the investigation, and whether the accused has been arrested and/or released on bail.
Based on this report, the magistrate makes a decision of whether there is sufficient ground for proceeding
against the accused. If he so rules, a formal chargesheet is prepared, detailing the offence with which the
accused is charged. If the magistrate feels that the police report does not provide sufficient grounds on
which to frame a charge, the case is dismissed at that point. The fraction of cases in which the police
report results in a formal chargesheet therefore reflects in part the quality of the investigation carried out
by the police; this will be a variable of interest in future work.12
We should note that our crime data comes from the very first stage in this process i.e. the filing of the
First Information Report. This requires two things: the victim or another informant must come to the
police station and give information regarding a crime, and the police officer must do his duty and make a
written record of this information. The victim in many cases can face physical or mental costs of coming
forward to file a FIR; in the case of vulnerable segments of society, s/he may even face the risk of
becoming the victim of a further crime if s/he reports it. Reporting is therefore likely to be lower for
11
All arrested persons must be produced before a magistrate and charged with a specific crime within twenty-four
hours or be released. This is the standard procedure under the habeas corpus requirements of Indian law.
12
We are in the process of putting together data on rates of chargesheeting in different crime categories over time.
certain types of crime than others, and also likely to be lower in places where the police force is less
motivated or more corrupt.13
3.2 Panchayati Raj: The Progress of Political Decentralization
Village-level governing bodies, popularly known as panchayats, have been known to exist in India since
ancient times. However, their real power, effectiveness and representativeness have varied considerably
over time. It is probably safe to say that women and members of the Scheduled Castes and Scheduled
Tribes were rarely represented in these bodies in pre-colonial times. Neither were these bodies
democratically elected. In the later period of British colonial rule, attempts were made to revive units of
local government. Some examples of such legislation include Lord Ripon’s resolution of 1882 to set up
district boards with elected officials (though suffrage was severely limited), and the report of the 1909
Royal Commission on Decentralization. There were strong disagreements on this issue during the framing
of independent India’s constitution, with Gandhi strongly supporting village democracy and B. R.
Ambedkar being sharply critical of them.14 The original constitution therefore presented a compromise
view: the setting up of village panchayats was recognized as a goal of the state, but not made mandatory.15
In 1957, a study team headed by Balvantray Mehta, a Member of Parliament, strongly recommended the
establishment of “statutory representative bodies” at the village level. Following this report, most states
enacted legislation to set up such village-level bodies, and by the 1960s, over 96% of inhabited villages
had panchayats of some sort. However, these panchayats were given very few resources to work with.
Further, over the next decade, several states failed to hold regular elections for these bodies, and many
new government programs bypassed these bodies in their implementation at the local level. By the late
1970s, it was acknowledged that the panchayat system of village government was either not working, or
working very poorly, in almost all Indian states.
A new era began in local government in 1978 when the state of West Bengal, followed by Karnataka,
took the lead in reforming their village level institutions along the lines of the recommendations of the
Asoka Mehta Committee (1978). This committee in fact recommended an explicitly political role for
panchayats, in the sense that they be democratically elected with political parties being allowed to contest
elections at village level. The states of West Bengal and Karnataka had also set up district-level
governance institutions as part of their “Panchayati Raj” reforms.
Following these successful experiments, a bill was introduced to provided constitutional protection for
such local bodies in 1989. However, several legislators and state governments opposed this bill, in part
because it was perceived as an attempt by the central government to bypass the state governments, and
because it attempted to impose a uniform structure on all areas without regard to their specific
circumstances. The bill failed to obtain enough votes in the upper house of parliament. A second version
of this bill was prepared by the next government in 1990, but never brought to debate. Finally, Narasimha
Rao’s Congress government introduced the 73rd Amendment Act in 1991. After some debate and
13
In some cases, there can be over-reporting of crimes. For instance, if there is a social stigma attached to premarital
sex, some instances of consensual sexual relations are registered as cases of rape.
14
It is clear that a principal part of B.R.Ambedkar’s criticism centered around the point that these village bodies
hardly ever represented the interests of the minority groups: “I hold that these village republics have been the
ruination of India. What is a village but a sink of localism, a den of ignorance, narrow-mindedness and
communalism?” (Remark in the Constituent Assembly, 4 November 1948).
15
Directive Principles of State Policy (Part IV of Indian Constitution): “The state should take steps to organize
village panchayats and endow them with such power and authority as may be necessary to enable to them function
as units of self-government.”
reference to a committee, the bill was finally passed in December 1992. Thereafter, it was ratified by a
majority of the states, and came into effect in April 1993.
This 73rd Amendment Act required each state to set up a three-tier system of local government,
comprising village, intermediate and district level governance bodies.16 All members of these local bodies
were to be directly elected by the people every five years, and the Act provided for the establishment of
State Election Commissions to conduct such elections. Twenty-nine areas of administration, including
decisions over health and education services, roads, sanitation and other local services were to be
devolved to these local government bodies (or Panchayati Raj institutions (PRIs) as they came to be
known). State Finance Commissions were to be set up to provide recommendations on revenue-sharing
and making grants to PRIs. The Act thus provided for a considerable degree of political, administrative
and fiscal decentralization to the local bodies.
Two further provisions were made to strengthen the representation of marginalized communities in these
local bodies. At least one-third of all council seats at the village, intermediate or district level were
required to be filled by women, and seats were also to be reserved for Scheduled Caste and Scheduled
Tribe communities in proportion with their population in that state or district. Further, the positions of the
heads of these local bodies were also to be reserved for women and members o the Scheduled Castes and
Scheduled Tribes in the same way. Similar provisions were made for urban local bodies as well.
How has this decentralization experiment been implemented in practice? The Constitutional amendment
required all states to amend existing laws or pass new laws in compliance with the 73rd Amendment. As
required, all states passed compliant legislation within one year.17 However, the actual provisions of the
Act were implemented very slowly. In terms of political decentralization, elections were eventually held
by all states to these local bodies, though in some states, they were delayed by litigation regarding the
provisions of the state legislation. In terms of administrative decentralization, as of 2007, only three major
states had operationally devolved all 29 functions to local governments. In terms of fiscal
decentralization, PRIs in most states were heavily dependent on grants from the state government for their
revenues.
In this paper, we focus on the changes in politician identity brought about by the reservation provisions in
the political decentralization aspects of the Panchayati Raj. We should note that this constitutional
amendment did not give local bodies any real control over the law and order machinery, and therefore is
unlikely to have an effect on crime through channels other than politician identity.
4. Data
4.1. Data on Crime
We obtained data on reported crime from various issues of the “Crime in India” publications of the
National Crime Records Bureau at the Ministry of Home Affairs. The Criminal Procedure Code of India
divides all crimes into two categories: (i) cognizable – which are dealt by the Police, in which a police
officer may arrest a person with or without a warrant and (ii) non-cognizable – which are generally left to
be pursued by the affected parties themselves in Courts. Only cognizable crimes are reported in these
16
Certain small states were required to set up only two-tier systems of local government.
The exceptions are Delhi, which has passed no Panchayati Raj legislation, Jammu & Kashmir which did not
explicitly pass legislation but amended the state laws to be in compliance, and the small states of Nagaland,
Mizoram and Meghalaya, which were not required to comply with this constitutional provision.
17
publications. There are two main types of cognizable crimes: those falling under the Indian Penal Code
(IPC), and those which are registered under the Special and Local Laws (SLL).
The Indian Penal Code (IPC) was first drafted in 1860, and records most common crimes under various
categories such as crimes against the body (murder, culpable homicide, hurt, negligence, kidnapping),
crimes against property (dacoity, robbery, burglary and theft), crimes against women (rape, dowry deaths,
cruelty by husbands or relatives, sexual harassment etc), crimes against public order (riots, arson),
economic crimes (criminal breach of trust, cheating counterfeiting) and other crimes. The exact
definitions of these crimes are listed in Appendix 1.
In addition to the IPC, crimes can also be recorded under a number of Special and Local Laws (SLL)
which have been enacted, for the most part, in the post-colonial period. These crimes are typically less
serious than the IPC crimes. For instance, dowry death is an IPC crime, while the giving or receiving of
dowry is an SLL crime; theft of an automobile is an IPC crime, while failure to register an automobile is a
SLL crime. Appendix 2 provides details of the SLL crimes recorded in our dataset. We should note that
several of these are also relevant mainly to women, such as the laws prohibiting prostitution, pornography
and the giving or receiving of dowry.
On a nationwide level, SLL crimes account for the majority of crimes reported in these data. Over the
period 1983-2007, SLL crimes are typically two-thirds of all crimes reported, while IPC crimes constitute
a third of the total recorded crimes. This is, of course, partly because IPC crimes are typically more
serious than SLL ones. In our analysis, we consider both types of crimes for completeness, but we do a
more detailed analysis of IPC crimes since they are typically the more serious ones and our framework
indicates that the effects of political identity changes are more likely to affect the serious crimes rather
than the non-serious ones.
We focus on analysis at the state level for now, although we do intend to pursue detailed district level
analysis in the near future. The advantage of using state level analysis is that we have access to data on a
much wider range of crime categories, i.e. not just IPC crimes, but SLL crimes as well. Further, many
control variables and additional outcomes are available only at state level. These include economic
variables, such as state level GDP levels and growth rates, as well as variables related to police strength
and staffing shortfalls, and additional outcomes such as the number of arrests and convictions. These
richer data facilitate better robustness checks and analysis.
Our main variable of interest is crimes against women. This consists of the following IPC crime
categories: rape, dowry deaths, sexual harassment, molestation, cruelty by husbands or relatives and
importation of women and girls, as well as the SLL provisions against prostitution, pornography, giving
and receiving dowry and sati. We will also examine the specific crime categories of rape and kidnapping
of women and girls.18 We should note that the reporting system for these crimes changes over time, as the
NCRB started reporting additional crime categories separately. For instance, the “crimes against women”
variable includes only rape in the period before 1995, but was expanded to include dowry death,
molestation, sexual harassment and cruelty by husband or relatives in 1995. A further category—
importation of women and girls—was included in 2001. The kidnapping of women and girls began to be
reported separately from total kidnappings in 1988. In all our regressions, we will therefore include year
fixed effects to control for such nationwide changes in reporting.
4.2. Measures of Political Decentralization
18
Under IPC classification, kidnapping of women and girls is usually listed under “crimes against the human body”
rather than crimes against women.
Our main measure of political decentralization is a dummy which equals one if marginalized sections of
society are given political representation. In the case of women, this dummy equals one in years following
the first local government election which implemented the “not less than one-third” reservation scheme
for women representatives. As Table 2 shows, the date of this first election varies considerably across the
major states of India.
There are three main reasons for the variation in election timing across states. First, several states already
had a system of local government even before the enactment of the 72nd Amendment. In most of these
cases, the state government waited for the term of office of incumbent local officials to expire before
conducting fresh elections in compliance with the 72nd Amendment. On the other hand, several states
chose to incorporate the provisions regarding women’s representation into their own state laws even
before the constitutional amendment came into effect. This was because they were aware of the
impending legislation due to the long process of passing this law (see Section 3.2) and had elections for
local bodies scheduled as per their existing system. For instance, West Bengal made major amendments
to their state-level legislation to provide reservation for women and SCs and STs in the 1993 election,
once the passage of the constitutional amendment was imminent. Kerala made a similar change to its law
in 1991. Other states already had reservation for women (Karnataka) or SCs and STs (Maharashtra and
Madhya Pradesh) long before the constitutional amendment. We can, of course, control explicitly for the
presence of a pre-scheduled local government election, but since all our regressions are with state fixed
effects, we expect this characteristic to be captured by the state fixed effect.
A second reason for variation in election timing is due to lawsuits challenging certain aspects of PR
implementation. For instance, elections in Bihar were delayed due to a lawsuit challenging the proposed
reservations for Other Backward Castes (OBCs) which had not been explicitly mandated by the
constitutional amendment. These can be regarded as reasonably exogenous factors in causing the delay.
A third reason is likely to be more endogenously determined: some states simply delayed the elections
due to various pretexts (the time is not right, lack of budget etc). Assam is one example of this. The state
had elections in 1992, and therefore should have had its first PR-compliant elections in 1997, but the state
government delayed on various pretexts, before finally conducting elections on December 2001. The
second round of elections was also delayed, taking place only in December 2007.
4.3 Empirical strategy
We conduct the analysis for the 17 major states of India over the period 1985-2007.19 Table 1 provides the
summary statistics for the crime data used in our analysis. Overall, we see total IPC crime per capita
declining after 1985, driven mostly by large declines in IPC crimes against property, against public order
and economic crimes. On the other hand, crimes against human body and against women show a rise over
this period on average. We should note that the big jump in crimes against women in 1995 is almost
19
These are the large states of India, which account for 97% of the total population and 98% of total crimes
reported: Andhra Pradesh, Assam, Bihar, Gujarat, Haryana, Himachal Pradesh, Jammu & Kashmir, Karnataka,
Kerala, Madhya Pradesh, Maharashtra, Orissa, Punjab, Rajasthan, Tamil Nadu, Uttar Pradesh and West Bengal.
Three new states—Chhattisgarh, Jharkhand and Uttarakhand—were carved out in 2001, from Madhya Pradesh,
Bihar and Uttar Pradesh respectively. All our analysis is conducted by merging these newly created states with their
original states, for the sake of comparability of sample units over time. All of these split states carried over the PR
legislation from their parent states, but this aggregation can sometimes cause measurement error in our dependent
variable: for instance, Bihar conducted local elections in 2001 and 2006, but Jharkhand has not conducted a single
PR election yet.
surely an artifact of the reporting of extra categories of crimes after 1994. However, rapes also exhibit a
steady rise over this period, but murders show a decline after 1995. SLL crimes against women show a
decline after 1990, but these are at a much lower level than the IPC crimes against women.
To gauge the impact of political representation, we run state-level regressions of crimes rates (per1000
population) on the decentralization variables defined above. Our base specification is the following:
(1)
ln Cst / Pst   s  t  Dst  X st   st
where C st is the crime rate for state s in year t, Pst is the population in state s and year t , αs is a fixed
effect for state s, βt is a fixed effect for year t, Dst  1 in years including and following the first election
with political representation for a given minority group, and X st is a set of state-time varying controls.
We have chosen to use the date of the first local election with reservation for women as our independent
variable to study the effects on crimes against women. All standard errors are clustered at the state level,
to account for possible correlated shocks to state-level crimes over time.
One concern with the above specification is that there can be many other factors which affect the rates of
crime against women in a given time and place. Literacy rates could affect awareness of victims’ legal
rights and influence reporting of crimes as well. Per capita income has also been found to be a factor in
higher rates of crime20. Crime is typically higher in urban areas relative to rural ones. There is some
anecdotal evidence that high male-female ratios result in increased rates of crime. Perhaps states which
implemented reforms earlier were those where the Chief Minister was a woman, and so we might
mistakenly attribute the effect of a higher-level woman representative to a lower-level one. We explicitly
control for all of these variables in our regressions. Finally, we also include measures of the size of the
state police force as a control for varying levels of commitment to crime deterrence across states, though
this might in fact be the channel through which political representation might impact crime.
A further concern is that the timing of policy changes may be endogenous to the trends in crime. In
addition to including state and time fixed effects, we also show specifications controlling for statespecific time trends. Further, we should note that several of the factors affecting the timing of this reform
(as detailed in section 3.2) can be considered exogenous to the trends in crime in those states. The fact
that we mainly observe the reform to be associated with an increase in reported crimes against women
also rules out incentives of state governments to prevent reporting of crimes in an attempt to make
themselves look good.
5. Political Representation and Reported Crimes against Women
5.1 Effects of Female Political Representatives on Crimes against Women
We find that political representation for women is associated with a large and significant increase in the
reported crimes against women. Table 3 shows the coefficients on the post-reform dummy Dst, when we
run the regression (1) for a range of crime categories and specifications. We report the results from
running regression (1) in Table 3. We find that reported crimes against women (per 1000 population) are
36% higher after political reservations for women are implemented (Column 1). This coefficient remains
robust when we control for a large number of demographic, economic and political controls (column 2),
20
Soares (2004) finds that a positive correlation between per capita income and reported crimes, which is explained
by larger reporting errors in less developed countries.
and even when we control for the police strength in the state at that time (column 3). This strongly
suggests that these effects are not due to a policy change in policing, but rather driven by the change in
the identity of the politician. Most of this increase is attributable to the IPC categories of crimes against
women rather than SLL ones.
A similar positive effect is also confirmed in specific types of crimes against women: there is a 19%
increase in rapes and a 13% increase in the kidnapping of women (estimates from Column 3). The results
are practically unchanged when we control for female literacy rather than overall literacy (Column 4).
These strong positive effects persist for most categories of crime against women even when we include
particularly strong controls, in the form of state-specific linear trends in addition to demographic controls
(Columns 5 and 6), though the coefficients are somewhat smaller in magnitude. We should note that this
specification controls for all linearly varying state-level variables, including any linear effect of
representation of women (and therefore might be “over-controlling” for the effects of the reform). As the
reader may recall, a positive effect of having a local level female politician in power on crimes against
women is most likely attributable to a net increase in the reporting of such crime.21
5.2 Effects on Other Categories of Crime
If the results in Table 3 are purely due to the change in the gender identity of the politician, then we
expect to see little or no effect on categories of crime where the victims are not women. One such
category is the kidnapping of men and boys, which is reported separately from 1988 onwards. Consistent
with our hypothesis, we do not see any signification relationship between this crime category and political
reservations for women (Table 4). Another implication of our framework was that we should observe
declines in reported crime for categories where the reporting probability is highest. Indeed, we find such a
negative coefficient for murder rates following the implementation of women’s political reservations.
However, this coefficient is not statistically significant, which could be because these are overall murder
rates, and not disaggregated by the gender of the victim. But the direction of the results are consistent
with our working hypothesis that the results of Table 3 are primarily driven by increases in reporting,
which are larger than any decreases in underlying crime rates.
We also show results for other crime categories which do not have any overt gender component: crimes
against property (robberies and burglaries), crimes against public order (riots and arson) and economic
crimes (counterfeiting, cheating and breach of trust). As we might expect, none of these crime categories
show any statistically significant relationship with the progress of women’s political reservations. This
makes us more confident that our results of Table 3 are indeed capturing the effect of women’s political
representation, rather than any other concurrent reforms which might have been implemented such as
changes to reporting rules, changes to policing strategy or overall economic growth or inequality trends.22
(see Ghosh, 2006, Prasad, 2008)
21
The above results are also robust to dropping states with some unusual characteristics: Jamu &Kashmir,
Karnataka (which was the first state to implement women's reservations) and Uttar Pradesh (the last state to
implement women's reservation). 22
For the impact of overall economic growth or rising inequality on crime in India, see Prasad (2008) and
Charmarbagwala and Sharma (2008).
5.3 Is the Probability of Action Taken Higher?
One of the underlying assumptions of our analysis here is that a match in the minority identity of the
politician and victim will lead to an increase in the probability of punitive action against those who
commit crimes against those minorities. We now examine the validity of this assumption using data on
arrests made on reported crimes. Our specification is similar to that in equation (1), except that the
dependent variables are now log of arrests per 1000 population and the log of ratio of arrests to reported
crimes. We also report preliminary results for crime-specific arrest rates, though we do not have this data
for all the years in our dataset as of now (hence the number of observations is much smaller).
We find that the number of arrests per 1000 population increases by a significant 19% after women’s
reservations are implemented (Table 5, Column 3 estimates). These estimates include controls for
economic and demographic variables, as well as for overall police strength as a proxy for the resources
available to the police. Arrests for rape also show a significant increase of over 18% during this period.
Further, we observe a significant increase in the arrests for kidnapping of women by as much as 21%,
while there is no such impact on arrests for kidnapping of males.
Finally, we also examined the results for the ratio of arrests to total crimes (table 6). For overall IPC, this
shows a significant rise after women’s empowerment. Unlike with the results for the number of arrests
reported in table 5, there is no significant effect on the arrest per crime for rape and kidnapping of
women. This is simply because the presence of female leaders results in not just the number of arrests,
but also the reporting of crimes itself. For instance, with rapes, there is a 19% increase in reported rapes
(table 3), as well as 19% increase in the number of arrests (table 5). Thus, focusing on the arrest per
crime variable alone, in the absence of crime reporting data, can yield a misleading picture.
5.4 Impact of Political Representation at Higher levels of Office in the State
We extend our analysis to consider the effects of having women leaders in political office at a higher level
of state government. Specifically, we consider members of the state legislative assembly (MLAs) as well
as the gender of the Chief Minister (highest officer) of the state. We should note that women are
significantly under-represented at these levels of government, where no mandated representation rules are
in place. In our data set, only 5.5% of the state legislators in any given year are female. Over this 22-year
time span, we find just 5 female Chief Ministers in India’s major states, accounting for about 8% of our
state-year observations. In this sense, we should note that the Panchayati Raj rules mandated a very large
increase in women’s representation over the existing status quo. This low level of representation means
that we are less likely to find any effect of state-level women representatives on rates of reported crime.
For women MLAs, we run a regression specification similar to (1), except that the independent variable
of interest is the fraction of female MLAs in power at the state level (Table 7, columns 1 and 2). Since
this fraction is potentially endogeneous, we instrument for the fraction of women MLAs using the fraction
of closely contested elections where the winner was female (Column 3).23 We should note that the OLS
and IV coefficients show a positive relationship between women’s representation and reported crime,
though none of these coefficients are statistically significant. This is most likely due to the low level of
overall representation.
23
A similar instrument is used by Clots-Figueras(2008) in her analysis of the effects of women legislators on public
good outcomes.
For the effects of a female Chief Minister, our specification follows equation (1) but uses a dummy
variable that equals 1 when the CM is female and 0 otherwise. In sharp contrast to the results on
representation at lower levels of government, we find that having a woman CM has a strong negative and
significant effect on reported crimes against women. We should note that the election of a woman to the
highest political office in the state is likely to be correlated with other trends in the state regarding the
status of women, and further, that there was no law which provided any specific exogenous variation in
the election of women Chief Ministers.
At this stage, we do not have strong micro-foundations to suggest specific channels through which female
leaders in higher levels of political office affect crime outcomes against women. Nevertheless, the
observed patterns at three different levels are consistent with the fact that the power to enforce punitive
action against criminals is strongest at the level of the Chief Minister (who is in charge of executive
decisions regarding law and order). At the other end, having proximate access to a local level female
politician can have a strong effect on women’s reporting of crimes – although she may have little power
to influence actual action against criminals. Therefore, one would expect the net reduction on actual
crimes to be stronger for female Chief Ministers and the net reporting of crimes to be greater with local
level women leaders. In ongoing work, we are currently exploring this and other hypotheses using district
level data on crime and political reservations.
6. Conclusions
This paper provides one of the first systematic analyses of the effects of mandated political representation
for minority groups on crime outcomes against them, using data from the Panchayati Raj experiment in
India. We find that having female leaders at the local government level induces strong positive and
significant effects on reporting of crimes by women – and no effects on any other categories of crime
where women are not victims. Interestingly, we find that having a woman as the head of state government
creates a strong reduction in the actual rates of crime against women. We conjecture that these
contrasting effects at the local versus state level are consistent with the stronger power for punitive action
vested in a the head of state – but also the greater voice available to minorities from access to leaders
from their own group at the local level.
In ongoing work, we plan to explore these effects further using district level data. In particular, we will
exploit the extra layer of representation provided to women through the requirement that one-third of
district council head positions must be filled by women. In most states, these one-third positions were
randomly allocated across districts in a given year, and rotated during the next election cycle, providing
plausibly random variation over time in the extent of women’s representation. We are also currently
pursuing similar analysis of the effects of political reservation for disadvantaged Scheduled Castes and
Scheduled Tribes under the same policy reform in India. A key difference for the latter analysis is that
these communities have already been provided with mandated reservation at the level of the state
legislature as well.
7. References
Bloch, Francis and Vijayendra Rao (2002) “Terror as a Bragaining Instrument: A Case Study of Dowry
Violence in Rural India,” American Economic Review, 92(4), pp. 1029-1043.
Charmarbagwala, Rubaina and Gunjan Sharma, 2008. “Crime and Inequality in India,” mimeo.
Clots-Figueras, Irma, 2008, “Women in Politics: Evidence from Indian States,” mimeo.
Cullen, Julie and Steven Levitt, 1999 “Crime, Urban Flight and the Consequences for Cities,” Review of
Economics and Statistics, Vol LXXXI (2) May 1999
Demombynes, Gabriel and Berk Ozler (2005) “Crime and Local Inequality in South Africa,” Journal of
Development Economics, 76, pp. 265-292.
Donohue III, John and Steven Levitt, 2001a. “The Impact of Race on Policing and Arrests”, Journal of
Economic Behavior and Organization, pp- 368-394
Donohue III, John and Steven Levitt, 2001b. “The Impact of Legalized Abortion on Crime.” Quarterly
Journal of Economics CXVI(2).
Dreze, Jean and Reetika Khera, 2000. “Crime, Gender and Society in India: Insights from Homicide
Data,” Population and Development Review, 26(2), pp. 335-352.
Ehrlich, Isaac, 1996 “Crime, Punishment and the Market for Offenses,” Journal of Economic
Perspectives, 10(1), pp. 43-67.
Fafchamps, Marcel and Bart Minten, 2006 “Crime, Transitory Poverty and Isolation: Evidence from
Madagascar,” Economic Development and Cultural Change, (54) pp. 579-603
Fajnzylber, Pablo, Daniel Lederman, Norman Loayza, 2002. “What causes violent crime?”, European
Economic Review 46(2002), 1323-1357.
Freeman, Richard, 1982 “Crime and the Labor Market”, NBER Working Paper 1031.
Freeman, Richard, 1994 “Crime and the Job Market,” NBER Working Paper 4910.
Glaeser, Edward 1999. “An Overview of Crime and Punishment,”
Ghosh, Arkadipta, 2006. “Electoral Cycles in Crime in a Developing Countries: Evidence from the Indian
States,” mimeo.
Gould, Eric, Bruce Weinberg and David Mustart, 1998 “Crime Rates and Local Labor Market
Opportunities in the US: 1979-1995”,
Government of India, 2008. “The State of Panchayats, 2007-08: An Independent Assessment,” Ministry
of Panchayati Raj, Government of India, New Delhi (www.panchayat.gov.in, accessed February 2009).
Grofman, Bernard and Lijphart, Arend, 1986. Electoral Laws and Their Political Consequences,
York: Agathon Press.
New
Institute of Social Sciences, 2000. Status of Panchayati Raj in the States and Union Territories of India,
2000, Institute of Social Sciences, New Delhi.
Kelly, Morgan 2000. “Inequality and Crime,” Review of Economics and Statistics, 82(4), pp. 530-539.
Levitt, Steven, 2004. “Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline
and Six that Do Not”, Journal of Economic Perspectives, 18(1), pp. 163-190.
Levitt, Steven, 1995. “The Effect of Prison Population Size on Crime Rates: Evidence from Prison
Overcrowding and Litigation”, NBER WP 5119.
Levitt, Steven, 1997. “Using Electoral Cycles in Police Hiring to Estimate the effect of Police on Crime,”
American Economic Review, 87(3), pp. 270-290.
Levitt, Steven, 1998, “The Relationship between Crime Reporting and Police: Implications for the Use of
Uniform Crime Reports,” Journal of Quantitative Criminology, 14(1), pp. 61.
Ludwig, J, G. Duncan and P. Hirschfield, 2001 “Urban Poverty and Juvenile Crime: Evidence from a
Randomized Housing Mobility Experiment,” Quarterly Journal of Economics 116(2), pp. 655-679.
Menon-Sen, Kalyani and A. K. Shiva Kumar (2001). "Women in India: How Free? How Equal?" United
Nations. http://www.un.org.in/wii.htm.
Miguel, Edward, 2005. “Poverty and Witch Killing,” Review of Economic Studies.
Miguel, Edward, Shanker Satyanath and Ernest Sergenti, 2004. “Economic Shocks and Civil Conflict: An
IV Approach”, Journal of Political Economy 112(4), pp. 725-753.
Mehlum, Halvor, Edward Miguel and Ragnar Torvik, 2006. “Poverty and Crime in 19th Century
Germany,” Journal of Urban Economics 59, pp 370-388.
Prasad, Kislaya, 2008, “Economic Liberalization and Violent Crime in India: 1985-2005 University of
Maryland”, mimeo.
Soares, Rodrigo, 2004. “Development, Crime and punishment: accounting for the international
differences in crime Rates,” Journal of Development Economics, 73, pp. 155-184.
Sotomayor, Sonia (2002) ʺRaising the Bar: Latino and Latina Presence in the Judiciary and the Struggle for Representation,ʺ Berkeley La Raza Law Journal, Spring symposium issue
Appendix 1: Definition of Major IPC Crimes
Crime
IPC Section #
Crimes against body
Culpable
299
homicide
Murder
Culpable
homicide not
amounting to
murder
300
Exception to
300
Causing death by
negligence
Attempt to murder
304A
Hurt
Kidnapping from
India
Kidnapping from
lawful
guardianship
Abduction
319
360
307
361
362
Definition
Causing death by doing an act with the intention of causing death, or with the intention of causing such bodily injury as is
likely to cause death, or with the knowledge that one is likely by such act to cause death.
Except in the cases hereinafter excepted, culpable homicide is murder if the act by which the death is caused is done with
intention of causing death or if it is done with the intention of causing such bodily injury as the offender knows to be likely to
cause the death of the person to whom the harm is caused.
Culpable homicide is not murder if: 1) The offender, whilst deprived of the power of self-control by grave and sudden
provocation, causes the death of the person who gave the provocation or causes the death of any other person by mistake or
accident; 2) The offender, in the exercise in good faith of the right of private defence of person or property, exceeds the power
given to him by law and causes the death of the person against whom he is exercising such right of defence without
premeditation, and without any intention of doing more harm than is necessary for the purpose of such defence; 3) If the
offender, being a public servant or aiding a public servant acting for the advancement of public justice exceeds the powers
given to him by law, and causes the death by doing an act which he, in good faith, believes to be lawful and necessary for the
due discharge of his duty as such public servant and without ill-will towards the person whose death is caused; 4) If it is
committed without premeditation in a sudden fight, in the heat of passion upon a sudden quarrel and without the offender’s
having taken undue advantage or acted in a cruel or unusual manner; 5) When the person whose death is caused, being above
the age of eighteen years, suffers death of takes the risk of death with his own consent.
Causing the death of any person by doing any rash or negligent act not amounting to culpable homicide.
Whoever does any act with such intention or knowledge, and under such circumstances that, if he by that act caused death, he
would be guilty of murder.
Causing bodily pain, disease or infirmity to any person.
Conveying any person beyond the limits of India without the consent of that person, or of some person legally authorized to
consent on behalf of that person.
Taking or enticing any minor under sixteen years of age if a male, or under eighteen years of age if a female, or any person of
unsound mind, out of the keeping of the lawful guardian of such minor or person of unsound mind, without the consent of such
guardian.
Compelling by force, or by any deceitful means inducing any person to go from any place.
Crimes against women
Rape
375
Dowry death
304B
Importation of girl
from foreign
country
Husband or
relative of
husband of a
woman subjecting
her to cruelty
366B
498A
Crimes against property
Theft
378
Extortion
383
Robbery
390
Dacoity
Preparation to
commit dacoity
391
399
A man is said to commit rape who, except in the case hereinafter excepted, has sexual intercourse with a woman under
circumstances falling under any of the six following descriptions: 1) Against her will; 2) Without her consent; 3) With her
consent, when her consent has been obtained by putting her or any person in whom she is interested in fear of death or of hurt;
4) With her consent, when the man knows that he is not her husband, and that her consent is given because she believes that he
is another man to whom she is or believes herself to be lawfully married; 5) With her consent, when, at the time of giving such
consent, by reason of unsoundness of mind or intoxication or the administration by him personally or through another of any
stupefying or unwholesome substance, she is unable to understand the nature and consequences of that to which she gives
consent; 6) With or without her consent, when she is under sixteen years of age.
Exception: Sexual intercourse by a man with his wife, the wife not being under fifteen years of age, is not rape.
Where the death of a woman is caused by any burns or bodily injury or occurs otherwise than under normal circumstances
within seven years of her marriage and it is shown that soon before her death she was subjected to cruelty or harassment by her
husband or any relative of her husband for, or in connection with, any demand for dowry, such death shall be called "dowry
death", and such husband or relative shall be deemed to have caused her death.
Importing into India from any country outside India or from the State of Jammu and Kashmir any girl under the age of twentyone years with intent that she may be, or knowing it to be likely that she will be, forced or seduced to illicit intercourse with
another person.
Any willful conduct which is of such nature as is likely to drive the woman to commit suicide or to cause grave injury or
danger to life, limb or health (whether mental or physical) of the woman; or harassment of the woman where such harassment
is with a view to coercing her or any person related to her to meet any unlawful demand for any property or valuable security is
on account of failures by her or any person related to her to meet such demand.
Taking dishonestly any moveable property out of the possession of any person without that person’s consent, or moving that
property in order to such taking.
Intentionally putting any person in fear of any injury to that person, or to any other, and thereby dishonestly inducing the
person so put in fear to deliver to any person any property or valuable security, of anything signed or sealed which may be
converted into a valuable security.
A special and aggravated form of either theft or extortion. The chief distinguishing element in robbery is the presence of
imminent fear of violence.
When five or more persons conjointly commit or attempt to commit a robbery.
Devising or arranging means necessary for the commission of dacoity.
Economic crimes
Criminal breach
of trust
405
Being in any manner entrusted with property, or with any dominion over property, and dishonestly misappropriating or
converting to one’s own use that property, or dishonestly using or disposing of that property in violation of any direction of law
prescribing the mode in which such trust is to be discharged, or of any legal contract, express or implied, which one has made
touching the discharge of such trust, or willfully suffering any other person to do so.
Cheating
415
By deceiving another person, fraudulently or dishonestly inducing the person so deceived to deliver any property to any person,
or to consent that any person shall retain any property, or intentionally inducing the person so deceived to do or omit to do
anything which he would not do or omit if he were not deceived, and which act or omission causes or is likely to cause damage
or harm to that person in body, mind, reputation or property.
Counterfeiting
231, 232, 489A Practicing deception, or knowing it to be likely that deception will thereby be practiced, by causing a piece of metal to appear
like Indian coin, or a false property mark, stamp or bank note to appear genuine.
Crimes against public order
Rioting
146
Whenever force or violence is used by an unlawful assembly, or by any member thereof, in prosecution of the common object
of such assembly, every member of such assembly is guilty of rioting.
Source. Indian Penal Code, 26 th Edition, 1987
Appendix 2: Definition of SLL crimes
Crime
Year of Act
Definition
Crimes against women
Immoral Traffic (Prevention) Act
1956
A prostitute who seduces or solicits shall be prosecuted. A client is guilty of consorting with prostitutes
and can be charged if he engages in sex acts with a sex worker within 200 yards of a public place or
"notified area". Babus or pimps or live-in lovers who live off a prostitute's earnings are guilty of a crime.
Any adult male living with a prostitute is assumed to be guilty unless he can prove otherwise.
Dowry Prohibition Act
1961
Indecent Representation of Women
(Probibition) Act
1986
Sati Prevention Act
1987
If any person, after the commencement of this Act, gives or takes or abets the giving or taking of dowry,
he shall be punishable.
No person shall publish, or cause to be published, or arrange or take part in the publication or exhibition
of, any advertisement which contains indecent representation of women in any form. No person shall
produce or cause to be produced, sell , let to hire, distribute, circulate or send by post any book,
pamphlet, paper, slide, film, writing, drawing, painting, photograph , representation or figure which
Whoever attempts to commit Sati and does any act towards such commission shall be punishable; if any
person commits Sati, whoever, abets the commission of such Sati, either directly or indirectly, shall be
punishable.
Crimes against SCs/STs
Protection of Civil Rights Act
1955
SC/ST Prevention of Atrocities Act
1989
Prohibits the practice of "untouchability" i.e. restricting access to public places, water sources, places of
worship, practice of any profession, occupation of residential premises etc on grounds of
"untouchability"
Prohibits a variety of atrocities towards SCs and STs, committed by non-SCs/STs, including stripping,
insulting, forced labor and sexual exploitation. Also penalizes the giving of false or fabricated evidence
aimed at implicating SC/ST members.
Other SLL crimes
Arms Act
1959
Narcotic Drugs & Psychotropic
Substances Act
1985
The Public Gambling Act
1867
The Central Excises & Salt Act
1944
No person shall acquire, have in his possession, or carry any firearm or ammunition unless he holds in
this behalf a licence issued in accordance with the provisions of this Act. No person shall manufacture,
sell, transfer, convert, repair, test or prove, or possess any firearm or any ammunition, unless he holds in
this behalf a licence issued in accordance with the provisions of this behalf a licence issued in accordance
with the provisions of this Act
No person shall cultivate any coca plant, opium poppy or any cannabis plant; or produce, manufacture,
possess, sell, purchase, transport, warehouse, use, consume, import or export any narcotic drug or
psychotropic substance.
Penalty for owning or keeping, or having charge of a gaming-house. Penalty for being found in gaminghouse. Penalty on persons arrested for giving false names and addresses.
Prohibits evasion of Excise duty or the removal of any excisable goods.
Explosives & Explosive Substances
Act
Indian Railways Act
1908
Registration of Foreigners Act
1939
Passport Act
1967
Essential Commodities Act
1955
Terrorist & Disruptive Activities Act
1987
Whoever conspires or attempts to commit, or advocates, abets, advises or incites or knowingly facilitates
the commission of, a terrorist act or any act preparatory to a terrorist act, shall be punishable with
imprisonment for a term which shall not be less than five years but which may extend to imprisonment
for life and shall also be liable to fine. Whoever harbours or conceals, or attempts to harbour or conceal,
any terrorist shall be punishable with imprisonment.
Antiquity & Art Treasure Act
1972
It shall not be lawful for any person to export and antiquity or art treasure.
1989
Prohibits causing of explosions, or making or manufacturing explosive substances with the intent of
endangering life or causing serious injury to property.
Fraudulently travelling or attempting to travel without proper pass or ticket. Needlessly interfering with
means of communication in a train. Obstructing railway servant in his duties. Other offenses related to
rail travel.
The Central Government may make rules for requiring any foreigner in India to report his presence to a
prescribed authority, and to provide other particulars to the relevant authority.
No person shall depart from, or attempt to depart from, India unless he holds in this behalf a valid
passport or travel document, or provide false information regarding the passport.
An Act to provide, in the interests of the general public, for the control of the production, supply and
distribution of, and trade and commerce, in certain commodities including cattle fodder, coal, component
parts and accessories of automobiles, cotton and woollen textiles, drugs, foodstuffs, iron and stell, paper,
petroleum, and any other commodity declared to be an essential commodity.
Table 1: Summary statistics for state-level crime variables (Major states, 1985 and later)
Panel A: Means for full data (1985-2007 for IPC and SLL; 1992-2007 for crimes against SC/ST)
#obs
Mean
S.D
Min
Max
Total IPC crimes per 1000 pop
391
1.855
0.658 0.448
3.301
IPC crimes against women per 1000 women
391
0.146
0.131 0.001
0.511
IPC crimes against human body per 1000 pop
391
0.253
0.194 0.022
0.817
IPC crimes against property per 1000 pop
391
0.446
0.191 0.099
1.087
IPC crimes against public order per 1000 pop
391
0.110
0.086 0.000
0.449
IPC economic crimes per 1000 pop
391
0.059
0.033 0.019
0.200
Murders per 1000 pop
391
0.035
0.018 0.011
0.150
Rapes per 1000 pop
391
0.015
0.010 0.000
0.049
Kidnapping of women and girls per 1000 women
340
0.038
0.031 0.005
0.149
Total SLL crimes per 1000 pop
SLL crimes against women per 1000 women
PCR crimes against SCs and STs per 1000 SC/ST pop
391
391
391
3.224
0.027
0.007
4.822
0.061
0.015
0.000
0.000
0.000
30.250
0.423
0.105
Total crimes against women per 1000 women
Total crimes against SCs per 1000 SC pop
Total crimes against STs per 1000 ST pop
391
271
225
0.173
0.161
0.088
0.137
0.167
0.124
0.001
0.000
0.000
0.570
0.766
1.031
Panel B: Means over time
1985
1990
1995
2000
2007
Total IPC crimes per 1000 pop
IPC crimes against women per 1000 women
IPC crimes against human body per 1000 pop
IPC crimes against property per 1000 pop
IPC crimes against public order per 1000 pop
IPC economic crimes per 1000 pop
Murders per 1000 pop
Rapes per 1000 pop
Kidnapping of women and girls per 1000 women
2.104
0.026
0.065
0.715
0.118
0.071
0.034
0.012
2.108
0.033
0.099
0.706
0.110
0.067
0.043
0.016
0.044
1.943
0.146
0.308
0.583
0.105
0.063
0.046
0.015
0.035
1.881
0.197
0.323
0.512
0.083
0.070
0.042
0.019
0.039
1.905
0.267
0.375
0.491
0.067
0.077
0.032
0.023
0.045
Total SLL crimes per 1000 pop
SLL crimes against women per 1000 women
PCR crimes against SCs and STs per 1000 SC/ST pop
3.451
0.021
0.000
3.955
0.039
0.020
2.739
0.019
0.012
0.852
0.020
0.006
1.927
0.014
0.005
Total crimes against women per 1000 women
Total crimes against SCs per 1000 SC pop
Total crimes against STs per 1000 ST pop
0.047
0.072
0.164
0.176
0.085
0.217
0.113
0.042
0.280
0.111
0.050
Table 2: Dates of Panchayati Raj Implementation Across States of India
Year of first election with
reservation for women
1987
1991
1992
1993
1994
1995
1996
2001
2002
2006
Total
Number of states
1
1
2
1
2
5
1
2
1
1
17
Table 3: Political Representation and Crimes against Women
Women's reservation implemented
Control for Control for stateControl for
statespecific time
female
specific
trends +
literacy time trends demographics
(4)
(5)
(6)
No controls
(1)
Demographic &
GDP controls
(2)
Police
controls
(3)
Total crimes against women
per 1000 women
Observations
R-squared
0.365*
(0.190)
391
0.85
0.365**
(0.149)
391
0.90
0.363**
(0.152)
391
0.90
0.360**
(0.152)
391
0.90
0.225**
(0.099)
391
0.95
0.225**
(0.087)
391
0.95
IPC crimes against women
per 1000 women
Observations
R-squared
0.276*
(0.148)
391
0.93
0.246*
(0.131)
391
0.94
0.240*
(0.132)
391
0.94
0.240*
(0.132)
391
0.94
0.272**
(0.126)
391
0.96
0.239*
(0.132)
391
0.97
SLL crimes against women
per 1000 women
Observations
R-squared
-0.019
(0.344)
379
0.77
0.001
(0.270)
379
0.81
-0.001
(0.274)
379
0.81
-0.005
(0.274)
379
0.81
-0.307
(0.273)
379
0.88
-0.203
(0.227)
379
0.90
Rapes per 1000 pop
Observations
R-squared
0.199**
(0.086)
391
0.81
0.199**
(0.076)
391
0.83
0.194**
(0.075)
391
0.83
0.192**
(0.076)
391
0.83
0.148*
(0.072)
391
0.91
0.101
(0.079)
391
0.92
Kidnapping of women
& girls per 1000 women
Observations
R-squared
0.160**
(0.056)
340
0.91
0.133**
(0.049)
340
0.92
0.134**
(0.050)
340
0.92
0.133**
(0.050)
340
0.92
0.135**
(0.059)
340
0.96
0.112*
(0.054)
340
0.96
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
State and Year FE
Clustering at state level
Demographic & economic controls
Control for woman CM
Control for police strength
Standard errors in parantheses, corrected for clustering at state-level.
*, ** and *** indicate significant at 10%, 5% and 1% respectively.
Regression for 17 major states and years 1985-2007. All crime variables are in logs.
Each cell represents the coefficient on the post-reform dummy for the outcome variables with
appropriate controls. Post-reform dummy equals 1 for years after the first local government elections
with reservations for women.
Demographic & economic controls include literacy, urbanization, female-male ratio, % population in farming and
real per capita state domestic product.
Table 4: Political Empowerment and Crimes not Targeted Towards Women
Women's reservation implemented
No controls
(1)
Demographic &
GDP controls
(2)
Police
controls
(3)
Kidnapping of men & boys
per 1000 men
Observations
R-squared
0.005
(0.135)
330
0.66
-0.003
(0.105)
330
0.68
0.023
(0.101)
330
0.69
Murders per 1000 pop
Observations
R-squared
-0.133
(0.098)
391
0.71
-0.132
(0.089)
391
0.73
-0.122
(0.082)
391
0.74
IPC crimes against property
per 1000 pop
Observations
R-squared
0.073
(0.073)
391
0.82
0.025
(0.051)
391
0.86
0.021
(0.052)
391
0.86
IPC crimes against public order
per 1000 pop
Observations
R-squared
0.193**
(0.085)
391
0.89
0.124
(0.075)
391
0.92
0.125
(0.074)
391
0.92
IPC economic crimes
per 1000 pop
Observations
R-squared
0.122
(0.075)
391
0.70
0.104
(0.062)
391
0.71
0.106
(0.063)
391
0.71
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
State and Year FE
Clustering at state level
Demographic & economic controls
Control for woman CM
Control for police strength
Standard errors in parantheses, corrected for clustering at state-level.
*, ** and *** indicate significant at 10%, 5% and 1% respectively.
Regression for 17 major states and years 1985-2007. All crime variables are in logs.
Each cell represents the coefficient on the post-reform dummy for the outcome variables with
appropriate controls. Post-reform dummy equals 1 for years after the first local government elections
with reservations for women.
Demographic & economic controls include literacy, urbanization, female-male ratio, % population
in farming and real per capita state domestic product.
Table 5: Political Representation and Arrest Rates
Women's reservation implemented
Demographic & GDP
No controls
controls
Police controls
(1)
(2)
(3)
Log (Arrests per 1000 population)
0.208***
(0.059)
391
0.70
0.186***
(0.050)
391
0.72
0.188***
(0.051)
391
0.72
Observations
R-squared
0.186**
(0.082)
221
0.85
0.184***
(0.063)
221
0.86
0.187**
(0.065)
221
0.86
Log (Arrests for kidnapping of women
per 1000 pop)
Observations
R-squared
0.231**
(0.080)
221
0.87
0.212***
(0.060)
221
0.88
0.217***
(0.063)
221
0.88
Log (Arrests for kidnapping of men
per 1000 pop)
Observations
R-squared
-0.175
(0.164)
212
0.74
-0.167
(0.123)
212
0.78
-0.141
(0.118)
212
0.79
State and Year FE
Clustering at state level
Demographic & economic controls
Control for woman CM
Control for police strength
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Observations
R-squared
Log (Arrests for rape per 1000 pop)
Standard errors in parantheses, corrected for clustering at state-level.
Regression for 17 major states and years 1985-2007. All crime variables are in logs.
Crime-specific arrest rates are for 1988-1991 and 1994-2002.
Demographic & economic controls include literacy, urbanization, female-male ratio, % population in farming and
real per capita state domestic product.
Table 6: Political Representation and Arrests per Crime
Women's reservation implemented
Demographic & GDP
No controls
controls
Police controls
(1)
(2)
(3)
Ratio of number of arrests to number of reported crimes (logs)
Total IPC
0.090**
(0.032)
Observations
391
R-squared
0.30
Rape
Observations
R-squared
Kidnapping of women
Observations
R-squared
Kidnapping of men
Observations
R-squared
State and Year FE
Clustering at state level
Demographic & economic controls
Control for woman CM
Control for police strength
0.094***
(0.030)
391
0.32
0.098***
(0.031)
391
0.32
-0.002
(0.031)
221
0.68
-0.001
(0.036)
221
0.69
0.004
(0.038)
221
0.69
0.027
(0.048)
221
0.69
0.051
(0.040)
221
0.71
0.058
(0.045)
221
0.71
0.036
(0.090)
211
0.64
-0.031
(0.094)
211
0.66
-0.027
(0.094)
211
0.66
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Standard errors in parantheses, corrected for clustering at state-level.
Regression for 17 major states and years 1985-2007. All crime variables are in logs.
Crime-specific arrest rates are for 1988-1991 and 1994-2002.
Demographic & economic controls include literacy, urbanization, female-male ratio, % population in farming and
real per capita state domestic product.
Table 7: Women Legislators and Crimes Against Women
Proportion of women legislators
Demograph
ic & GDP
controls
Police controls
IV
(1)
(2)
(3)
Woman Chief Minister
Demographic
& GDP
Police
controls
controls
(4)
(5)
Total crimes against women
per 1000 women
Observations
R-squared
1.858
(1.983)
391
0.89
2.536
(2.000)
391
0.89
5.206*
(2.827)
391
0.89
-0.285*
(0.149)
391
0.89
-0.286*
(0.151)
391
0.89
IPC crimes against women
per 1000 women
Observations
R-squared
-0.871
(1.148)
391
0.94
-0.318
(1.068)
391
0.94
1.546
(1.678)
391
0.94
-0.266**
(0.117)
391
0.94
-0.268**
(0.119)
391
0.94
SLL crimes against women
per 1000 women
Observations
R-squared
1.214
(3.328)
379
0.81
0.970
(3.482)
379
0.81
6.800
(6.551)
379
0.81
0.142
(0.236)
379
0.81
0.141
(0.236)
379
0.81
Rapes per 1000 pop
Observations
R-squared
0.515
(1.328)
391
0.82
0.968
(1.200)
391
0.82
1.015
(1.697)
391
0.82
-0.198***
(0.054)
391
0.82
-0.199***
(0.054)
391
0.82
Kidnapping of women
& girls per 1000 women
Observations
R-squared
0.509
(1.137)
340
0.91
1.299
(0.995)
340
0.92
1.155
(2.083)
340
0.92
-0.289***
(0.075)
340
0.92
-0.289***
(0.075)
340
0.92
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
State and Year FE
Clustering at state level
Demographic & economic con
Control for woman CM
Control for police strength
Y
Standard errors in parantheses, corrected for clustering at state-level.
*, ** and *** indicate significant at 10%, 5% and 1% respectively.
Regression for 17 major states and years 1985-2007. All crime variables are in logs.
IV regression using the proportion of women legislators winning in close elections (less than 5% vote margin)
as an instrument for the overall proportion of women legislators.
Demographic & economic controls include literacy, urbanization, female-male ratio, % population
in farming and real per capita state domestic product.