Bail Decision Making in Karnataka’s Lower Criminal Courts: Why the Nature of the Offences Matters


– Sudhir Krishnaswamy and Deekshitha Ganesan

The rate of under-trial detention in India, at 67% of the total prison population (NCRB Prison Statistics, 2014-16), is a serious challenge to the effectiveness of the criminal justice system. Despite the regular intervention of the Supreme Court, most recently through In Re Inhuman Conditions in 1382 Prisons ((2016) 10 SCC 17), the consistently high level of under-trial detention remains a pressing concern that calls for our urgent and collective attention.

Reform efforts have traditionally focused on conditions of detention and the performance of institutional mechanisms designed to control under-trial detention, such as Under-Trial Review Committees and provision of legal aid (CHRI, 2016; Amnesty International, 2017). Academic writing on bail decision making has also primarily focused on a doctrinal analysis of bail provisions in the Code of Criminal Procedure, 1973 (CrPC). However, these studies fail to explain the reason for the high under-trial detention rate in India – what brings so many individuals into the prison system?

CLPR’s study: Pre-trial bail decision making in trial courts

The Centre for Law and Policy Research’s empirical study on bail decision making at the pre-trial stage attempts to answer this question. We rely on in person court observations of first production cases over 6 weeks and pre-trial court records in three districts in Karnataka – Bengaluru, Tumakuru and Dharwad – to evaluate the influence of multiple substantive and procedural factors on the bail decision making process. A better understanding of the bail decision making process in the trial court is essential to ensure that under-trial detention is at the optimal level necessary for an effective criminal justice system.

At a bail hearing, three outcomes are possible. First, the accused may be released on bail. Second, the court may remand the accused to police custody for further investigation. Third, the court may direct the accused to judicial custody in the prison system. But, what drives a court to grant or reject bail? Does it depend on the offence or the availability of quality legal representation? Alternatively, does the interaction between the accused and the court determine whether bail is granted?

In this post, we explore the influence of one such substantive legal factor – nature of the offence – and focus on how it shapes the bail decision making process. Hypothetically, bail is less likely to be granted where a person is accused of a serious offence which may cause harm to society.

Bail Outcomes at First Production

An accused is first produced before the trial court within 24 hours of being arrested. ‘First production’ is the first crucial point of intervention of courts in the criminal justice process, where the court makes a decision on whether to grant bail or detain the accused in police or judicial custody.

We observed 378 first production cases in Bangalore (284), Dharwad (54) and Tumakuru (40). We begin with a broad review of the outcomes at first production in each district.

Figure a: Outcome of First Productions

Judicial custody was the most common outcome at first production (Figure a). 74% of first productions in Bengaluru resulted in judicial custody, with Dharwad and Tumakuru recording 67% and 57%, respectively. Police custody was granted in fewer cases – only in 1 in 5 cases in Bengaluru and 1 in 50 cases in Dharwad. Notably, in Bengaluru, the number of police custody orders was more than three times that of orders granting bail.

We note a wide variation among the three districts with respect to bail orders. Bail was granted in about 25% of the cases in Dharwad and Tumakuru but in only 6% cases in Bengaluru. This variation could be due to differences in police, prosecutorial and judicial cultures in the three districts. Data on these counts was not readily available through the court observations. However, we can measure the influence of other legal factors which might contribute to these differences, such as the nature/seriousness of the offence.

Relationship between Bail Outcomes and the Nature of the Offence

Criminal offences can be broadly classified as offences under the Indian Penal Code (IPC) and offences under special and local laws (SLLs). ‘SLLs’ is an umbrella term for a range of non-IPC criminal statutes enacted to tackle a specific social issue through penal sanctions such as laws to protect children; laws to control production, supply and distribution of narcotics; or terrorism offences.

Offences are also categorised into bailable (Section 436) and non-bailable offences (Section 437) under the CrPC. While bail must be granted as a matter of right to persons accused of bailable offences, police and courts have discretion to grant bail in case of non-bailable offences.

Our study measured the influence of these two characteristics which could determine the nature/seriousness of an offence: (a) the statutory basis i.e. whether the offence originates from the IPC or SLLs and (b) its classification as bailable or non-bailable in Schedule – I of the CrPC.

(a)    Statutory Basis of Offence

Across the three districts, SLL offences were most commonly registered under the Protection of Children from Sexual Offences Act, 2012 and the Narcotic and Psychotropic Substances Act, 1985.

Bengaluru240 (84.5%)44 (15.5%)
Dharwad36 (66.7%)18 (33.3%)
Tumakuru21 (52.5%)19 (47.5%)

Table 1: Number of IPC Cases and SLL Cases

Although the proportion of IPC and SLL offences varied across the three districts, we note that the number of IPC offences outnumbers SLL offences in all three districts. In fact, IPC offences in Bengaluru were 5 times the proportion of SLL offences (Table 1).

Irrespective of the statutory basis of the offence, judicial custody was the most likely outcome of a bail decision at first production. Police custody was more likely to be ordered in case of SLL offences than IPC offences in Bengaluru and Dharwad, where accused persons were twice and six times as likely to be remanded to police custody for SLL than IPC offences. In fact, police custody was not ordered in any IPC offence in Dharwad. In Tumakuru, however, the likelihood of police custody being ordered were similar for IPC (19%) and SLL offences (21%).

In all three districts, bail was more frequently granted in IPC offences than SLL offences. Accused persons were three times as likely to secure bail for an IPC offence when compared to an SLL offence.

Figure b: Statutory Basis of Offences and Outcomes of First Production

One reason behind this is that SLL statutes often prescribe special procedures for arrest and place restrictions on bail, which significantly impacts bail outcomes. For instance, Section 37 of the Narcotic and Psychotropic Substances Act, 1985 provides that certain offences, including offences involving commercial quantities, are non-bailable.

(b)   Classification as Bailable v. Non-Bailable

When we undertook this study, we expected only a few bailable cases at first production as police bail is routinely granted for bailable offences. This expectation was confirmed in all three districts. Although 20% of the first production cases in Dharwad were bailable in nature, the overall proportion of bailable offences was lesser than non-bailable offences in all three districts.

Figure c: Outcome of first productions – Bailable v. Non-Bailable

In Bengaluru (48%) and Dharwad (82%), the court granted bail in more bailable offences than non-bailable offences. However, bail was not granted by the court to any person accused of a bailable offence in Tumakuru.

27% non-bailable offences were granted bail in Tumakuru, which appears to be the outlier. On the other hand, a significantly smaller proportion of non-bailable offences received bail in Bengaluru (2%) and Dharwad (12%).

Does the nature of the offence shape bail decision making?

From the data set, we note that the nature of the offence appears to have a significant influence on a bail outcome at first production. Offences which are perceived by the court as less serious are more likely to receive bail, irrespective of whether we consider the statutory basis of the offence or its classification in Schedule – I, CrPC.

The common trend in all three districts was that bail was granted at higher rates for IPC offences when compared to SLL offences. Significant variations in the outcome on first production are visible across the three districts depending on whether the offence committed was under the IPC or an SLL. Further, bail was granted more frequently for bailable than non-bailable offences. While Tumakuru was the exception, it is important to bear in mind that only 3 bailable cases were before the courts.

Although the classification of offences as bailable and non-bailable in Schedule – I, CrPC is often understood to be based on the seriousness of offences, it requires revision. For example, Section 324 of the IPC which punishes voluntarily causing hurt by dangerous weapons with imprisonment of three years and fine, is a non-bailable offence. However, voluntarily causing grievous hurt, which is punishable with imprisonment of seven years and fine, is classified as a bailable offence.


The data on the relationship between the nature of offences and bail outcomes presents two interesting results.

Despite common statutory texts, investigative procedures, administrative structures, and a jurisdictional high court, there was significant variation in outcomes in the three districts. This requires careful attention, as geography or the location of the offence also appears to influence bail decision making.

Further, the direct relationship between nature of the offence and bail outcomes reveals that judges are largely following the applicable legislations. However, some penal statutes are over-prescriptive, which contributes to the high rate of under-trial detention. Therefore, the impetus for reform is at two levels – to streamline the bail decision making process and amend existing laws.

Besides nature of the offence, other substantive and procedural factors may also significantly shape bail decision making in lower criminal courts. For instance, the availability of effective legal representation could improve the likelihood of securing bail, as could greater interaction between the court and the accused. We present a similar analysis of the influence of other substantive and procedural factors in our forthcoming report on bail decision making in the pre-trial stage in Karnataka. More information on this report may be accessed herehere and here.

[1] The most commonly invoked SLLs invoked in the 3 districts are:

  • Bangalore – Narcotic and Psychotropic Substances Act, 1985; Immoral Traffic (Prevention) Act, 1956; Arms Act, 1959.
  • Dharwad – Protection of Children from Sexual Offences Act, 2012; Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act, 1989; Negotiable Instruments Act, 1881.
  • Tumakuru – Protection of Children from Sexual Offences Act, 2012; Immoral Traffic (Prevention) Act, 1956; Karnataka Police Act, 1963.

For instance, the POCSO does not restrict the powers of the court to grant bail and extends the provisions of the CrPC on bail to offences under it. However, the NDPS Act designates particular offences under it as non-bailable and places fetters on the powers of the court to grant bail.

Sudhir Krishnaswamy is a co-founder of the Centre for Law and Policy Research and a partner at Ashira Law. He is also currently a faculty member at the Azim Premji University.  A graduate of the National Law School of India University, Bangalore, he has formerly served as the Dr. B.R. Ambedkar Visiting Professor of Indian Constitutional Law at Columbia Law School.

Deekshitha Ganesan is a research associate at the Center for Law and Policy Research and an alumna of the National Law School of India University, Bangalore. She has also worked with Khaitan & Co., Mumbai with their Litigation and Dispute Resolution team.


Please enter your comment!
Please enter your name here