Why some Delhi High Court judges are less likely to grant bail than others

Whether a person accused of a crime waits for trial at home or in prison often depends on one decision: bail. The Supreme Court’s position for decades has been that bail is the rule and jail is the exception. However, the law allows judges wide room to decide these individual cases of liberty. A new study that looks at 4,459 bail cases decided by the Delhi High Court asks whether a judge’s background is linked to how they use their discretion to grant bail.

The study, published in the Journal of Law and Courts, a Cambridge University Press journal, analyses 4,459 bail applications decided by 34 judges of the Delhi High Court between 2023 and 2024. It found that judges promoted from the district judiciary granted bail at lower rates than judges appointed from among practising lawyers, even after accounting for differences in the cases they heard. The authors describe this as “a robust descriptive association between professional pathway and bail outcomes.”

Simply put: a trial court judge who is subsequently elevated to the High Court is less likely to grant bail than a High Court judge, who was appointed directly from the Bar.

Currently, 40% of judges serving in the Delhi High Court are from the district judiciary. But the findings of the study matter beyond Delhi since every High Court draws judges from both routes, and their standing is equal as judges of a constitutional court. They could also go on to become judges of the Supreme Court. Article 217 of the Constitution sets the eligibility criteria for an individual to be appointed as a judge of a High Court: ten years at the Bar or ten years in judicial office. Conventionally, at least one-third of the sanctioned strength of the High Court is reserved for service judges. The number of serving judges from the district judiciary or the Bar varies at any given point in time depending on judges’ tenures and vacancies.

Why does it matter which judge hears a bail plea?

Nearly 70% of Indian prisons are filled with undertrials. Given the pace of trials, bail is the most crucial element in determining whether the process becomes the punishment for an accused.

A bail plea before the High Court is often filed after the trial court has rejected bail. In our three-tier legal system, trial courts are bound by rather strict statutory requirements and tighter processes, while High Courts, as constitutional courts, have greater room for discretion.

Speaking in 2024, then Chief Justice D.Y. Chandrachud said that trial court judges often choose to play it safe by not granting bail in criminal cases. For such applicants, the High Court bench is where liberty is decided. “People who should be getting bail in trial courts are not receiving it,” he said, forcing them to approach the High Courts.

What has the study found?

Story continues below this ad

Of the 4,459 bail applications, 72.5 per cent were granted. Bar judges decided 3,102 applications and granted bail in 80.3 per cent. Service judges decided 1,357 and granted bail in 54.8 per cent. The raw gap is 25.5 percentage points.

After adjusting for case features such as bail type, whether the victim died, whether an earlier application had been rejected, the number of legal sections and whether a senior advocate appeared, the paper estimates that service judges were 23.7 percentage points less likely to grant bail.

If the law is the same, how can outcomes differ so much?

India doesn’t have one bail law, but bail provisions in different statutes set different standards. Judges have significant discretion to account for the particular facts of each case. Courts weigh the nature of the accusation, the likely punishment, the risk of tampering with evidence and the strength of the case, but no formula says how much weight each carries. It also depends on the accused, whether one is considered a flight risk.

Story continues below this ad

The paper treats this flexibility as the reason differences can appear: “If bail were formulaic, professional background should be irrelevant; the discretionary latitude inherent in bail adjudication is what makes pathway-linked differences plausible.” In other words, where the law leaves more to judgment, the background of the person exercising that judgment has more room to show.

A plausible explanation is that trial court judges who spend most of their careers in district courts will likely carry their practices to high courts and give greater weight to the state’s case against an accused.

Advocate Bharat Chugh said it is difficult to shake off a habit built over a long career. “It’s difficult to let go of that hard-wiring that happens over a very long career of being slightly conservative. You become a HC judge if you’re a trial court judge, maybe 30 years after you join the service; you cannot suddenly become a different person. That is a possible hypothesis for the huge gap in the dataset.”

Chugh also pointed to what Bar judges bring with them. “When you practice for several years and represent individual clients, standing for their personal liberty in courts, then maybe you are slightly more aware of the very serious consequences that being in jail for even one day has for somebody’s life. You have sat with clients and families and seen families uprooted because somebody was arrested and not given bail.”

Do such biases matter?

Story continues below this ad

Case outcomes have a direct impact on the lives of litigants. A 2023 study by Nitin Kumar Bharti and Sutanuka Roy in the Journal of Public Economics (2023) studied 668 judges deciding 323,380 bail cases in Uttar Pradesh and found that judges exposed to Hindu-Muslim riots in early childhood were about 16 per cent more likely to deny bail than the average judge. The study ruled out religious favouritism as the explanation but linked the effect of riots to judges’ relative trust in individuals versus state institutions.

If biases are apparent in an individual judgment, then they can often be rectified through an appeal. However, systemic inequities can be ascertained through large data sets. For example, a bench of judges comprising both those elevated from the Bar and those from the district judiciary could help address bias in bail decisions.