The Many Afterlives of Najeeb – The Supreme Court’s Reference on Delay as a Ground for Bail under the UAPA – Constitutional Law and Philosophy

[This is a guest post by Rushil Batra.]

This year, and more particularly the last few weeks, have been important for understanding the Supreme Court’s bail jurisprudence under the Unlawful Activities (Prevention) Act, 1967 (“UAPA”). The immediate reason is that coordinate benches of the Supreme Court have taken sharply different approaches to the question of whether delay in trial can, by itself, justify the grant of bail in UAPA cases (see here and here). That disagreement has now resulted in a reference to a larger bench.

The problem is not merely that different benches have reached different outcomes in individual bail matters (see here). It is a well-settled position of law that the ratio of a case is based on the facts before the concerned Court. The deeper problem here is that the Court has treated a number of such “facts” in Najeeb as relevant to how it understands the role of delay in trial as a ground for bail, and therefore to whether delay can be the sole ground for granting bail. Is delay in trial an independent constitutional ground for bail under Article 21? Or is it merely one factor to be weighed alongside the seriousness of allegations, the statutory restrictions under Section 43D(5), the stage of trial, and national security concerns?

This essay attempts to provide some answers to this question by offering a typology of the Court’s jurisprudence, and showing how greatly it varies. While the reference is necessary given these widely different interpretations, it must specifically clarify two issues i.e., first, whether delay is an independent ground for bail; and second, if it is, the more vexed question of “how long is too long” for delay to transgress constitutionally permissible limits and justify the grant of bail.

The Statutory Problem: Why Najeeb Became Important

Before turning to the conflicting post-Najeeb cases, it is important to identify what Najeeb itself held. The case arose under the UAPA, where bail is ordinarily governed by the stringent restriction under Section 43D(5). By the time the matter reached the Supreme Court, the accused had already spent more than five years in custody, the trial had barely progressed, and there was no real likelihood of it being completed within a reasonable time.

It was in this context that the Supreme Court held that the statutory restrictions on bail under special laws cannot completely eclipse the jurisdiction of constitutional courts. It held that while courts must ordinarily respect the legislative policy against bail in special statutes like the UAPA, that policy cannot justify indefinite pre-trial incarceration. The Court therefore held that the rigours of such provisions would “melt down” where there is no likelihood of the trial being completed within a reasonable time, and continued custody would violate Article 21 [¶18].

The principle emerging from Najeeb, therefore, is relatively clear that delay in trial can operate as an independent constitutional ground for bail, distinct from the statutory inquiry under Section 43D(5). The question is not merely whether the allegations are prima facie true. The question is whether continued incarceration, despite the accused still being an undertrial, has become constitutionally impermissible.

However, Najeeb does not fully clarify what happens to the factual matrix of the case once delay is established. In other words, it remains unclear whether prolonged incarceration makes the merits of the case entirely irrelevant, or whether the seriousness of the allegations, the nature of the material against the accused, and the stage of trial continue to have some residual relevance. Put differently, does Article 21 operate as an independent route to bail once delay becomes constitutionally excessive, or must the Court still balance delay against the merits of the prosecution’s case under Section 43D(5)? This ambiguity lies at the heart of the conflicting approaches that have emerged after Najeeb.

To be clear, Najeeb was not the first case in which the Supreme Court granted bail in a UAPA matter on account of prolonged incarceration. Even earlier, in cases such as Angela Harish Sontakke v State of Maharashtra, the Court had recognised that the seriousness of allegations must be balanced against the period already undergone in custody and the likely time within which the trial may be completed.

I. The First Life: Attempts to Narrow Najeeb

The first line of cases after Najeeb attempts to narrow its reach. The most prominent example is Gurwinder Singh v. State of Punjab. In that case, the accused had spent nearly five years in custody, which was close to the period of incarceration in Najeeb. Yet, the Court refused to grant bail, holding that the reliance placed on Najeeb lacked “depth and substance” on the facts of the case [¶32].

The Court distinguished K.A. Najeeb on three principal grounds [¶32]. First, it noted that in Najeeb, all co-accused had already been tried and sentenced to terms not exceeding eight years. Since the accused in Najeeb had already undergone more than five years in custody, the Court treated the substantial portion of the likely sentence already served as a relevant factor. Second, the Court emphasised that in Najeeb, the trial was unlikely to conclude in the near future because it had been severed from the trial of the co-accused, and a large number of witnesses were yet to be examined. By contrast, in Gurwinder Singh, the trial was already underway, and twenty-two witnesses, including protected witnesses, had been examined. Third, the Court relied on the gravity of the allegations, including alleged links to a banned terrorist organisation and large financial transfers requiring further investigation, and concluded that releasing the accused posed a serious risk of influencing witnesses.

On this basis, the Court held that the delay in Gurwinder Singh was not comparable to the delay in Najeeb. The troubling part of Gurwinder Singh, however, was not simply that bail was denied. The concern was the Court’s statement that “mere delay” in trials involving grave offences cannot be used as a ground to grant bail [¶32]. If read broadly, this proposition significantly weakens Najeeb and suggests that delay is not an independent ground for bail, but only one amongst many factors to be weighed against the seriousness of the allegations. After all, most UAPA cases will involve grave allegations! If the seriousness of the offence can itself defeat a delay-based bail claim, then the constitutional protection recognised in Najeeb becomes uncertain at the very point where it is most needed.

To be sure, later benches attempted to confine Gurwinder Singh to its facts. In the cases that followed, the Court distinguished Gurwinder Singh in several decisions, signalling its discomfort with the judgment and recognising that its holding was either confined to its own facts or inconsistent with the broader constitutional reasoning underlying Najeeb. For instance, in Shoma Kanti Sen v. State of Maharashtra, while granting bail to an accused who had been in custody for over six years, the Court referred to Gurwinder Singh and held:

He [ASG Mr. Nataraj] cited the case of Gurwinder Singh (supra) in which the judgement of K. A. Najeeb (supra) was distinguished on facts and a judgment of the High Court rejecting the prayer for bail of the appellant was upheld…But this was a judgment in the given facts of that case and did not dislocate the axis of reasoning on constitutional ground enunciated in the case of Najeeb (supra). [¶39]

Similarly, in S.K. Javed Iqbal v. State of Uttar Pradesh, while granting bail to an accused person who had been incarcerated for nine years, the Court once again reiterated that Gurwinder Singh must be confined to its own facts:

…[I]n Gurwinder Singh, the trial was already underway and that twenty two witnesses including the protected witnesses have been examined. It was in that context, the two Judge Bench of this Court in Gurwinder Singh observed that mere delay in trial pertaining to grave offences cannot be used as a ground to grant bail. [¶31].

Just when it might have appeared that Gurwinder Singh was an outlier, the Supreme Court’s decision in CBI v. Dayamoy Mahato, authored by Justice Karol and Justice Kotishwar Singh, suggested otherwise. In that case, the accused had undergone over twelve years of custody. The High Court had granted bail solely on the ground of prolonged incarceration and delay in trial. However, while reversing the High Court’s order, the Supreme Court described its approach as “fallacious” and held that it had erred in releasing the accused by “merely” invoking Article 21 precedents[¶12]. The Court further observed that the individual accused cannot “always be the centre of attention,” particularly in cases involving offences against the State and concerns of national security [¶16]. The striking feature of the decision is that the Court subordinated even an extreme period of custody to considerations of national security, without clearly explaining why twelve years of undertrial incarceration did not warrant constitutional intervention.

The more recent decision in Gulfisha Fatima v. State also belongs to this line of cases. While the Court granted bail to some co-accused, it denied bail to Umar Khalid and Sharjeel Imam. In doing so, it expressly rejected an automatic or time-based application of Najeeb, holding that delay does not operate as a “trump card” that automatically displaces statutory restraint [¶56]. Instead, the Court treated delay only as a trigger for heightened scrutiny, to be balanced against the gravity of the offence, the role attributed to the accused, the strength of the prima facie case, and the proportionality of continued incarceration.

Thus, at the very least, judgments like Gurwinder Singh, Dayamoy Mahato, and Gulfisha Fatima show that the constitutional logic articulated in Najeeb is neither uncontested nor settled within the Court’s bail jurisprudence under the UAPA. Rather than treating delay as an independent constitutional ground for bail, these cases fold delay back into a broader balancing exercise, where the merits of the prosecution’s case and the seriousness of the allegations continue to play a decisive role.

II. The Second Life of Najeeb: Delay Acknowledged But Bail Denied

The second line of cases accepts Najeeb in principle, but avoids granting bail in result. In these cases, the Court acknowledges that prolonged incarceration may violate Article 21, but instead of treating delay as a ground for immediate release, it provides “managerial remedies” such as expediting the trial, fixing timelines, or allowing the accused to renew the bail application later.

This approach is visible in Union of India v. Saleem Khan. By the time the matter reached the Supreme Court, the accused had spent more than five and a half years in custody, charges had not been framed, and the trial had not commenced. Here, more than one hundred witnesses were therefore yet to be examined. The Court recognised that an accused cannot be left to languish indefinitely without the prospect of a fair and speedy trial. Yet, it did not grant bail. Instead, it directed that the trial be completed within two years.

A similar pattern appears in Harpreet Singh Talwar v. State of Gujarat. In this case, the accused had been in custody for nearly three years. The Court acknowledged that, in an appropriate case, the rigour of Section 43D(5) must yield to Article 21. But it held that such relaxation is “not automatic” [¶24].  It denied bail, while granting liberty to the accused to apply again after six months or after substantial progress in trial.

This approach appears moderate, but it is actually doctrinally unstable. If prolonged incarceration has already crossed the point at which it becomes constitutionally impermissible, then a direction to expedite trial does not answer the Article 21 violation, and merely postpones relief. More importantly, directions to expedite trials often do not work in practice. UAPA trials involve voluminous records, protected witnesses, multiple accused, and long witness lists. A judicial direction to conclude the trial within a fixed period may sound reassuring, but it often fails to address the structural reasons for delay.

The effect is that the accused remains in custody, the trial continues slowly, and Najeeb becomes a deferred promise. This line of cases therefore recognises the constitutional problem but responds to it through managerial remedies rather than liberty-based remedies.

III. The Third Life: Applying Najeeb Faithfully

The third line of cases applies Najeeb more directly. Here, the Court treats delay in trial as an independent constitutional ground for bail and refuses to allow the seriousness of allegations to become a complete answer to Article 21.

The clearest example is S.K. Javed Iqbal v. State of Uttar Pradesh. The accused had spent over nine years in custody, and the State argued that the trial should be expedited instead of bail being granted. The Supreme Court rejected this argument, holding that “when a trial gets prolonged, it is not open to the prosecution to oppose bail of the accused-undertrial on the ground that the charges are very serious” [¶22]. In this case, the accused was also facing prosecution under Section 16 of the UAPA, which carries a maximum sentence of life imprisonment. Yet, the Court did not treat the seriousness of the charge, or the possibility of life imprisonment, as sufficient to defeat the claim for bail. Instead, it held that even under stringent penal statutes, courts must “lean in favour of constitutionalism and the rule of law of which liberty is an intrinsic part” [¶32].

A similar approach can be seen in Javed Gulam Nabi Sheikh v. State of Maharashtra, where the accused had undergone around four years of incarceration, charges had not been framed, and the State was unable to provide any clear timeline for completion of trial. In these circumstances, the Court held that if the State cannot protect the accused’s fundamental right to a speedy trial, it should not oppose bail merely by pointing to the seriousness of the offence. The Court expressly stated that “Article 21 of the Constitution applies irrespective of the nature of the crime” [¶18].

This is the strongest version of Najeeb. It understands the right to speedy trial not as a weak equitable consideration, but as a constitutional limit on pre-trial incarceration. On this view, the statutory bar under Section 43D(5) cannot authorise indefinite detention. Nor can the seriousness of allegations convert undertrial custody into a form of punishment before conviction.

However, even within this line of cases, there remains uncertainty. Bail was granted after nine years in Javed Iqbal and after around four years in Javed Gulam Nabi Sheikh. While both cases apply Najeeb faithfully, they do not provide a clear standard for when delay becomes constitutionally excessive. The Court has still not answered the central question: how long is too long?

The Unresolved Question: How Long is Too Long?

This question arises only if one first accepts that delay in trial can operate as an independent ground for bail. While Najeeb appears to support that reading, the Court’s later jurisprudence has made the position far less clear.

Even in decisions that take liberty seriously, the Court has been reluctant to treat the mere passage of time as automatically entitling an accused to bail. For instance, in Syed Iftikhar Andrabi, the Court observed that “no reading of Najeeb suggests that the mere passage of time, divorced from all surrounding circumstances, mechanically entitles an accused to release.” [¶30]. (The case has already been analysed by Abhinav Sekhri here).

In that sense, the real difficulty after Najeeb is not simply whether delay matters. Almost all benches now accept, at least formally, that it does. The real disagreement is over when delay becomes constitutionally unacceptable, and what legal consequence must follow once that point is reached.

In Tapas Kumar Palit v. State of Chhattisgarh, the Court came close to articulating a temporal benchmark. It observed that if an accused receives a final verdict only after six to seven years of undertrial incarceration, then the right to speedy trial under Article 21 stands infringed. The Court noted:

We may sound as if laying some guidelines, but time has come to consider this issue of delay and bail in its true and proper perspective. If an accused is to get a final verdict after incarceration of six to seven years in jail as an undertrial prisoner, then, definitely, it could be said that his right to have a speedy trial under Article 21 of the Constitution has been infringed… (emphasis mine) [¶14].

This is one of the few moments in the post-Najeeb jurisprudence where the Court appears to gesture towards a concrete standard. But the Court’s own docket complicates this. In some cases, bail has been granted after four years of custody. In others, it has been denied even after twelve years.  

This produces the central difficulty in the post-Najeeb line of cases. If delay is not an independent ground for bail, then the “how long is too long” question may never truly arise, because delay will always be balanced against seriousness, stage of trial, and the merits of the prosecution’s case. But if delay is an independent constitutional ground, then the Court must say when prolonged incarceration crosses the line from permissible pre-trial custody into unconstitutional punishment. That is the question the larger bench must now answer.

Conclusion: What Should the Larger Bench Clarify?

The larger bench must clarify, first, whether Najeeb recognises delay in trial as an independent constitutional ground for bail. If it does, then the Najeeb inquiry cannot be collapsed into Section 43D(5). The statutory question is whether the accusation is prima facie true. The constitutional question is whether continued undertrial incarceration has become incompatible with Article 21.

Second, the Court must clarify what role the facts of the case continue to play once delay is established. Do the seriousness of allegations, the stage of trial, the nature of evidence, and national security concerns remain relevant? Or does excessive delay itself justify bail? This ambiguity has allowed later benches to read Najeeb in sharply different ways.

Third, the Court should be careful not to treat expeditious-trial directions as a substitute for bail. Where charges have not been framed, hundreds of witnesses remain, or the State cannot give a realistic timeline, such directions merely postpone the Article 21 problem.

Finally, if delay is an independent ground, the Court must answer the harder question: how long is too long? Any benchmark should operate as an upper limit, not a lower threshold. It should not mean that bail must be denied before that period is crossed. But once crossed, the burden on the State to justify continued incarceration must become extremely heavy.

Ultimately, the reference is about whether Najeeb is a constitutional rule or only a discretionary exception. The larger bench must decide which understanding (or afterlife) survives. It will be intriguing to see how Justice Surya Kant, as CJI, will constitute a special bench to interpret a judgment he had himself authored in 2021!