Last verified: 13 May 2026

A cold morning in August 2024. Outside Tihar’s main gate, a small knot of lawyers waited with a discharge file in hand. Inside the Supreme Court that same morning, a two-judge bench read out a judgment that would reshape how trial courts treat Section 45 of the Prevention of Money Laundering Act, 2002. After 17 months in custody without a single witness being examined, a former Deputy Chief Minister of a Union Territory walked out on bail.

For seven years after Parliament resurrected it through the Finance Act, 2018, the law of PMLA Section 45 bail had felt close to absolute. The provision demanded two things before a court could even consider bail: the public prosecutor had to be heard, and the court had to be satisfied that there were reasonable grounds to believe the accused was not guilty and was unlikely to commit any offence on bail. That second prong, applied at the bail stage, looked like a near-impossible threshold. Most trial courts treated it that way too.

Understanding PMLA Section 45 bail today means understanding how that arc bent across nine landmark Supreme Court rulings between 2017 and 2026. The provision was struck down once in November 2017. Parliament revived it four months later. A three-judge bench upheld the revival in July 2022. Then, across August and September 2024, the same Court softened it through four rulings in quick succession. In February 2025, it sharpened again when a different bench refused to let the twin conditions be applied “casually”. And in January 2026, the Court added a new variable: prosecutorial wherewithal, the practical capacity of the agency to actually finish the trial.

What this means for a practitioner drafting a bail application today is uncomfortable. You can’t lean on one case. You have to read the arc. You have to know which judgment helps which prong, when the proviso opens and when it shuts, and what the latest Supreme Court mood signals for the trial court you’re appearing before. Before walking through that arc, the basics.

Section 45 of the Prevention of Money Laundering Act, 2002 imposes twin conditions for bail: the public prosecutor must be given an opportunity to oppose the application, and the court must be satisfied that there are reasonable grounds for believing the accused is not guilty and is unlikely to commit any offence on bail.

---

That paragraph is the heart of the bar. Everything else in this post is how courts have read those two prongs across a quarter-century of litigation, and how a practitioner builds a bail application that actually clears them.

---

---

What does Section 45 of the PMLA actually say?

Most practitioners reach for Section 45 only after the Enforcement Directorate has already filed a remand application. By then, the wording feels secondary; what matters is what the trial court will accept. That’s a mistake. The text of the section is the floor every argument has to clear, and the sub-sections beyond 45(1) carry tactical weight that practitioners regularly under-use.

Section 45 sits within the framework of the Prevention of Money Laundering Act, 2002 and governs every bail application brought by a person accused of an offence under the Act. It has three operative sub-sections, plus a first proviso and a second proviso. Sub-section (1) sets the twin conditions. Sub-section (1A) carries an under-cited police-investigation authorisation. Sub-section (2) makes bail conditions a mandatory part of the order itself.

Section 45(1): the twin conditions verbatim

The operative language reads that no person accused of an offence under the Act shall be released on bail unless the public prosecutor has been given an opportunity to oppose the application, and where the prosecutor opposes, the court must be satisfied that there are reasonable grounds for believing that the accused is not guilty of such offence and that they are not likely to commit any offence while on bail. Both prongs are cumulative. Both have to be met.

“Reasonable grounds to believe” is not “beyond reasonable doubt”, and that distinction matters. At the bail stage, the court doesn’t decide guilt. It examines the prosecution material on its face. In practice, courts have read this as something close to a prima facie standard, though Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929 insisted the threshold is meaningful rather than nominal. What experienced practitioners know is that the prosecutor’s opposition is the easy prong; the second prong, the not-likely-to-commit limb, is where most applications actually fail.

Section 45(1A): police investigation authorisation

Sub-section (1A) authorises any officer of the Enforcement Directorate to investigate offences under the Act. It rarely surfaces in bail arguments because it isn’t a bail provision; it’s an authorisation clause. But its placement inside Section 45 has procedural consequences. When defence counsel challenges the legality of an ED investigation (jurisdiction, designation of the investigating officer, transfer between zonal offices), 45(1A) is the textual hook. The Tarsem Lal court referenced this sub-section while dealing with the cognisance question, and it is frequently overlooked in general overviews of Section 45.

Section 45(2): bail conditions as part of the order

Sub-section (2) provides that the limitation on grant of bail under sub-section (1) is in addition to the limitations under Section 480 of the Bharatiya Nagarik Suraksha Sanhita, 2023 (the bail framework that was previously governed by Section 437 of the CrPC). What that means in practice: once the court is satisfied on the twin conditions, it still has to impose appropriate bail conditions under the BNSS framework, and those conditions are part of the order. Practitioners who draft a “naked” bail prayer without including a conditions menu (passport surrender, reporting, residence) are leaving the trial court to write its own terms, and the trial court usually writes harsher terms than the defence would have proposed.

A common question at this stage is what “reasonable grounds” actually requires the court to do. The Supreme Court has clarified that the bail court has to look at the material on record, not pretend it doesn’t exist. But the bail court doesn’t try the case. The right framing in the application is: “the prosecution material, even taken at face value, does not establish a prima facie nexus between the applicant and proceeds of crime”. That sentence does more work than three pages of denial.

The pitfall most counsel fall into is reading Section 45 as if it were a self-contained code. It isn’t. It layers on top of the general bail framework under Section 480 BNSS and Section 483 of the Bharatiya Nagarik Suraksha Sanhita, 2023 (the Sessions Court and High Court bail provisions formerly numbered Section 437 and Section 439 of the CrPC). Treat Section 45 as an additional filter, not a replacement, and the architecture of the bail application becomes clearer.

Is money laundering bailable or non-bailable under the PMLA?

This is the first question every accused person’s family asks, usually within hours of an Enforcement Directorate arrest. The short answer? Non-bailable, but the label is only the start of the analysis.

Offences under the PMLA are cognisable and non-bailable. That classification flows from the First Schedule to the BNSS (and earlier from the First Schedule to the CrPC) read with Section 45 PMLA itself. Cognisable means the police, or in PMLA cases the ED, can arrest without a warrant. Non-bailable means there is no automatic right to bail; the accused has to apply, and the court has discretion. The “non-bailable” label tells you the door is closed. It doesn’t tell you the door is locked.

Why the “non-bailable” tag is only the starting point

Every non-bailable offence under Indian criminal law is, in theory, bailable on application. Bail in non-bailable offences is governed by Section 480 of the Bharatiya Nagarik Suraksha Sanhita, 2023 for the Magistrate’s Court and Section 483 BNSS for the Sessions Court and High Court. Those provisions give the court discretion. For most non-bailable offences (say, a Section 302 IPC homicide or a serious BNS offence), the court weighs flight risk, evidence tampering, the seriousness of the offence, and the antecedents of the accused. The triple test. The traditional bail jurisprudence.

PMLA changes the test. Section 45 layers the twin conditions on top of that traditional framework. So a money-laundering bail court is doing two things at once: applying the general triple-test analysis under Section 480 / Section 483 BNSS, and additionally satisfying itself on the twin conditions under Section 45 PMLA. Both filters must clear. That’s why “non-bailable” understates the bar; the real picture is “non-bailable plus a statutory restriction with two cumulative prongs”.

How Section 45 layers on top of Section 437/439 CrPC and Section 480/483 BNSS

With the Bharatiya Nagarik Suraksha Sanhita, 2023 having replaced the CrPC from 1 July 2024, the bail-application title now reads “Application under Section 45 PMLA read with Section 483 BNSS” for an application before the Sessions Court or the High Court, and “read with Section 480 BNSS” for one before the Magistrate. Most practitioner templates still carry the older CrPC numbering. That isn’t fatal, but it dates the application and signals to the bench that the lawyer hasn’t updated. A small detail. Not a small signal.

What experienced practitioners know is that the layered architecture is also an opportunity. The triple-test analysis under Section 483 BNSS still happens. So defence counsel can argue both that the twin conditions are satisfied, and independently that the traditional bail factors (no flight risk, deep roots in the community, willingness to surrender passport) tilt in favour of release. The two arguments reinforce each other. Treating them as one collapses the analysis and weakens the prayer.

Want to argue and win complex bail matters under the PMLA, NDPS Act, and UAPA? LawSikho’s Diploma in Criminal Litigation and Trial Advocacy trains you in evidence-led bail advocacy, twin-conditions argumentation, and case-law mapping across stringent statutes, with live moot-court bail hearings and section-wise drafting templates.

How the twin conditions evolved: a 2002 to 2026 doctrinal arc

The cleanest way to understand PMLA Section 45 bail is chronologically. The provision didn’t always read the way it reads now, and the constitutional gloss on it has changed direction at least four times in nine years. Practitioners who treat it as a static text get blindsided.

The arc has five phases. Each phase shifts what the bail court is doing, and each phase produced one or two landmark rulings that subsequent benches now anchor to.

Phase	Period	Statutory state	Doctrinal posture
Phase 1	2002 to Nov 2017	Original Section 45(1): twin conditions tied to scheduled offences carrying more than three years	Strict bail bar; few constitutional challenges succeed
Phase 2	Nov 2017 to Mar 2018	Section 45(1) struck down by Nikesh Tarachand Shah v. Union of India, (2018) 11 SCC 1; provision inoperative	Twin conditions effectively suspended; bail flows under general BNSS framework
Phase 3	Mar 2018 to Jul 2022	Finance Act, 2018 amends Section 45(1); trigger shifts to “offences under this Act”	Twin conditions revived; constitutional challenge pending
Phase 4	Jul 2022 to mid-2024	Vijay Madanlal Choudhary v. Union of India (2022) upholds the revived Section 45	“Heinous crime” framing; high bail bar firmly back
Phase 5	Aug 2024 onwards	Post-2024 quartet softens the application; Kanhaiya Prasad (2025) rebalances; Jan 2026 wherewithal ruling adds a new prong	Contextual application; speedy-trial doctrine reads liberty back into Section 45

2002 to 2017: the original twin conditions and their reach

When the Section 45 of the Prevention of Money Laundering Act, 2002 provision was first enacted, the twin conditions applied only to those scheduled offences (the “Part A” offences) that carried imprisonment of more than three years. That triggering language matters because it tied the bail bar to the predicate offence, not to the laundering charge itself. A person accused of laundering proceeds from a less serious predicate offence theoretically faced a lower bail threshold. The architecture was clumsy but constitutionally defensible.

November 2017: the first death of Section 45(1)

In Nikesh Tarachand Shah, a two-judge bench struck down Section 45(1) as unconstitutional. The Court held that linking the bail bar to the predicate-offence punishment produced arbitrary results: two persons accused of the same money-laundering offence could face wholly different bail tests depending on which scheduled offence underpinned the charge. That arbitrariness, the bench held, violated Articles 14 and 21 of the Constitution. For four-and-a-half months, Section 45(1) was inoperative; this also reopened the broader question of the Act’s constitutional validity, which the courts continued to test on related provisions.

March 2018: the Finance Act revival

Parliament responded through the Finance Act, 2018. Section 45(1) was redrafted: the trigger was now “any offence under this Act”, not the predicate offence. The arbitrariness vice identified by Nikesh Tarachand Shah was, in the Government’s reading, cured. Every PMLA accused now faced the same bail test, regardless of which scheduled offence underlay the charge. Whether the amendment actually fixed the underlying constitutional concern was a separate question, and one that would take four more years to reach the Supreme Court.

July 2022: Vijay Madanlal Choudhary and the constitutional re-blessing

In July 2022, a three-judge bench in Vijay Madanlal Choudhary upheld the 2018 amendment. The 545-page judgment described money laundering as a heinous crime with grave socio-economic consequences and held that the revived Section 45, read in light of the Act’s purpose, did not offend Articles 14 or 21. The twin conditions were back, this time with a Supreme Court imprimatur.

Looked at across these phases, the historical pattern is unmistakable: Parliament and the Court have repeatedly oscillated between liberty and deterrence on Section 45, and the post-2024 phase is the most pronounced swing yet. Practitioners who write applications based on a single phase’s case law inevitably mis-pitch the argument.

PMLA Section 45 bail: the 2017-2026 doctrinal arc

Seven milestones that re-shaped the twin conditions in nine years

1. Nov 2017
   Nikesh Tarachand Shah
   Twin conditions struck down under Articles 14 and 21.
2. Mar 2018
   Finance Act 2018 amendment
   Twin conditions revived; trigger shifted to “offences under this Act”.
3. Jul 2022
   Vijay Madanlal Choudhary
   Revived Section 45 upheld; “heinous crime” framing locked in.
4. May 2024
   Tarsem Lal
   Twin conditions inapplicable when accused appears on summons post-cognisance.
5. Aug-Sep 2024
   Sisodia / Kavitha / Prem Prakash / Senthil Balaji
   Speedy-trial doctrine read into Section 45; foundational-facts burden re-balanced.
6. Feb 2025
   Kanhaiya Prasad
   Pendulum back: twin conditions remain mandatory; bail not granted “casually”.
7. Jan 2026
   Arvind Dham wherewithal test
   If prosecution lacks wherewithal to finish trial, Section 45 cannot prolong detention.

Source: Supreme Court of India rulings, 2017-2026. Read in sequence, not isolation.

Nikesh Tarachand Shah (2017): why the original Section 45 was struck down

The 2017 strikedown is foundational. Every subsequent challenge to Section 45 has to engage with what Nikesh actually held, and a surprising number of bail applications cite it for propositions it never decided.

The Articles 14 and 21 reasoning

The two-judge bench in Nikesh Tarachand Shah v. Union of India, (2018) 11 SCC 1 approached Section 45 through the equality lens of Article 14 first. The Court asked: does the original twin-conditions provision treat similarly situated persons similarly? Its answer was no. The bail bar applied to those accused of laundering proceeds of a scheduled offence carrying more than three years’ imprisonment, but not to those accused of laundering proceeds of a less-serious predicate. Two persons charged with the identical PMLA offence faced different bail tests because of an upstream variable that had nothing to do with the laundering itself.

That, the Court held, was manifest arbitrariness. The Article 21 layer followed: a procedurally arbitrary bail bar that did not bear a rational nexus to the offence being prosecuted could not survive the procedural-fairness requirement that Article 21 demands. The provision wasn’t merely impractical. It was unconstitutional on the text.

The “manifest arbitrariness” prong

Manifest arbitrariness, as a doctrinal tool, was relatively fresh in 2017. The Court applied it surgically to Section 45(1) without attempting a broader constitutional overhaul. That meant the surrounding provisions of the PMLA (Section 19 arrest, Section 50 statements, Section 24 reverse onus) survived. Only the twin-conditions sub-section, and only insofar as it incorporated the predicate-offence trigger, fell. Practitioners reading the case for support on PMLA’s broader constitutional infirmities will find limited material; the holding is narrower than it first appears.

What survived the strikedown

After Nikesh, three things remained intact. First, the non-bailable classification under the BNSS / CrPC schedule. Second, the prosecutor’s right to be heard at bail stage under the general framework. Third, the standard discretionary bail jurisprudence under what is now Section 483 BNSS. For roughly four months, that’s how PMLA bail applications were argued, and the High Courts that decided applications in that window did so on triple-test reasoning, not twin-conditions reasoning. A small number of those orders remain instructive on how to argue the not-likely-to-commit prong even today.

What practitioners often miss is that Nikesh did not say twin conditions are inherently unconstitutional. It said the specific 2002 architecture was. Parliament could redraft the provision, and Parliament did. Citing Nikesh today as a general anti-twin-conditions authority is a recurring error; counsel who advance this proposition are quickly steered back to Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929, which is now the controlling authority on the constitutional question.

The 2018 Finance Act amendment: how Parliament revived Section 45

The Government’s response to Nikesh was unusually swift. The Finance Bill of 2018 carried an amendment to Section 45(1) PMLA, and once enacted, the twin conditions were back. Most practitioners know this happened. Fewer can articulate what exactly changed in the text, and why the redrafted version survived where the 2002 version did not.

The drafting move: from “scheduled offences carrying >3 years” to “offences under this Act”

The original Section 45(1) opened with a condition tied to the predicate: “Notwithstanding anything contained in the [CrPC]… no person accused of an offence punishable for a term of imprisonment of more than three years under Part A of the Schedule shall be released on bail unless…” That language is what Nikesh found arbitrary. The trigger sat outside the laundering offence itself.

The 2018 amendment rewrote the opening to read “no person accused of an offence under this Act shall be released on bail unless…”. The trigger was now internal: it referred to the PMLA offence, not the predicate. Every person accused under the Act now faced the identical bail test. The Article 14 vice was cured at the textual level.

Why the amendment survived where the original failed

The amendment had two effects. First, it removed the cross-offence arbitrariness that Nikesh had identified. Second, and less obviously, it expanded the reach of the twin conditions: anyone accused under the PMLA, regardless of how serious the underlying predicate, now faced the bar. That second effect was a hardening, not a softening. The bail threshold went up across the board.

What experienced practitioners know is that the amendment is also where the bail mathematics fundamentally shifted. Before 2018, a PMLA accused whose predicate carried a maximum of two years’ imprisonment faced general bail jurisprudence. After 2018, the same accused faced the twin conditions. For an accused of money laundering with a smaller predicate, the 2018 amendment made bail materially harder, not easier. That distinction is often overlooked in general summaries of the amendment, and it matters in practice. The 2022 Vijay Madanlal ruling tacitly approved this expansion. The bench treated it as cured arbitrariness rather than expanded reach.

Vijay Madanlal Choudhary (2022): the revival blessed

For four years between 2018 and 2022, defence counsel argued that the revived Section 45 carried over the same constitutional flaws as the 2002 version. The Supreme Court took the question up in Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929 and resolved it in July 2022. The judgment is long, the reasoning is dense, and the consequences are still being worked out by trial courts.

The three-judge bench, the 545-page judgment, and the “heinous crime” framing

The bench, of three judges, framed money laundering as a uniquely serious offence with cross-border, socio-economic, and national-security implications. That framing did real work. Once the offence was characterised as heinous, the constitutional weighing tilted against a liberty-prioritising reading of Section 45. The Court held that the twin conditions, viewed against the backdrop of the gravity of the offence, were not unconstitutional.

The judgment also dealt with a string of other PMLA challenges: Section 19 arrest powers, Section 50 statements, the ECIR requirement, the reverse-onus presumption under Section 24. All survived, with varying degrees of judicial gloss. But the bail holding is what changed the day-to-day reality for PMLA accused. After Vijay Madanlal, trial courts treated the twin conditions as fully binding, and bail-grant rates fell sharply between mid-2022 and mid-2024.

How the Court read the 2018 amendment as curing the Nikesh defects

The Court’s logic on the Article 14 issue was straightforward: the amended Section 45 applies uniformly to all PMLA accused; the arbitrariness identified in Nikesh has been removed by Parliament’s textual redrafting. On the Article 21 issue, the bench held that the procedural fairness requirement is satisfied when the bail court applies the twin conditions in light of the statutory scheme and the gravity of the offence. There is no constitutional right to a relaxed bail standard in offences of this gravity. Article 21, on this reading, is not violated when a stringent statutory bar imposes a real but rational threshold.

The unresolved questions Vijay Madanlal left open

What the bench did not address, and what later benches have had to fill in, were the practical questions. How long can the twin conditions justify continued detention when the trial has not started? Does the first proviso operate mandatorily or discretionarily? Do the conditions apply when the accused has not been arrested and appears on summons? Each of those gaps became a separate Supreme Court ruling in 2023 and 2024.

In practice, the Vijay Madanlal moment was the high-water mark of the strict-application phase. Within two years, the Court itself would start softening the application, and the softening would come from benches that did not formally disturb the 2022 holding. The case is still good law on the constitutional question. It’s no longer the complete picture on how the twin conditions actually operate.

The post-2024 dilution: how the Supreme Court re-balanced the twin conditions

Between May and September 2024, four Supreme Court rulings in quick succession changed how the twin conditions are applied. None of them formally overruled the 2022 Vijay Madanlal Choudhary judgment. All of them softened it in practice. Practitioners now refer to this stretch as the post-2024 quartet, and the cases work together more than separately. Reading them in isolation produces a misleading picture.

Speedy trial read into Section 45

In August 2024, a two-judge bench decided the bail application of a former Deputy Chief Minister of a Union Territory who had spent 17 months in custody in Manish Sisodia v. Directorate of Enforcement, 2024 INSC 595. The bench held that the right to a speedy trial, anchored in Article 21, must be read into Section 45 PMLA and into Section 439 of the CrPC (now Section 483 BNSS). Prolonged incarceration without trial progress, where the trial is unlikely to conclude soon, tips the constitutional balance back towards liberty. The accused walked out the same day.

The ruling does not say twin conditions never apply. It says they cannot operate as a back-door indefinite detention when the prosecution cannot move the trial forward. That distinction is critical. Trial courts that read the order as a general bail relaxation get reversed; trial courts that apply it as a delay-specific carve-out get affirmed.

Bail is the rule returns to PMLA

Days after the August 2024 ruling, the same bench decided the bail application of a senior aide of a sitting Chief Minister in Prem Prakash v. Union of India through Directorate of Enforcement, 2024 INSC 637. The judgment did something Vijay Madanlal had not: it imported the foundational principle that “bail is the rule, jail is the exception” expressly into PMLA jurisprudence. The Court held that the Enforcement Directorate must establish prima facie foundational facts of money laundering before the twin conditions even bite. Without that threshold showing, the not-likely-to-commit prong cannot do all the work.

A common question practitioners raise is whether this means the prosecutor’s case is being tested at the bail stage. The answer, on a careful reading, is no. The Court is not tying the bail court to a guilt-or-innocence determination. It is asking the bail court to check whether the prosecution has even reached the starting line, whether there is a discernible nexus between the accused, the predicate offence, and the proceeds of crime. That’s a prima facie filter, not a trial. Most defence counsel under-use this gateway.

Section 45(1)(ii) cannot authorise indefinite detention

In September 2024, in V. Senthil Balaji v. Deputy Director, Directorate of Enforcement, 2024 INSC 739, the Court widened the speedy-trial principle. The bench held that Section 45(1)(ii) (the not-likely-to-commit prong) cannot be read as authorising indefinite pre-trial detention. Where the trial is not likely to conclude within a reasonable time, constitutional courts must intervene. The accused, a former Cabinet Minister of a southern State, had been in custody for an extended period without trial commencement. The bench granted bail with stringent conditions.

What experienced practitioners know is that the Senthil Balaji ruling supplies the legal hook for a delay argument even where the absolute length of custody is shorter than in Sisodia. The framing is about prospective trial timeline, not just retrospective custody length. A nine-month custody case with no realistic trial date in the next eighteen months can lean on Senthil Balaji; counsel who insist on a 17-month threshold are reading the case too narrowly.

How trial courts are reading the post-2024 quartet

Trial courts have begun to treat the quartet as a four-part test that supplements the Vijay Madanlal twin conditions, rather than displaces them. The questions trial courts now ask are: has the ED established prima facie foundational facts (Prem Prakash); how long has the accused been in custody (Sisodia); is the trial realistically going to conclude (Senthil Balaji); and, where applicable, has the accused even been arrested or are they appearing on summons (Tarsem Lal, discussed in the next section). Where one or more of these factors tip towards liberty, courts have been more willing to find that the twin conditions are met on the available record.

The pitfall to flag here is that the post-2024 quartet is sometimes read as a defence-side victory. It isn’t, not entirely. The same courts that granted bail also imposed exacting conditions (passport surrender, daily reporting, no contact with witnesses, residence-in-jurisdiction). And the very next year, the Court would push back through Union of India v. Kanhaiya Prasad, 2025 INSC 210, reminding trial courts that the twin conditions remain mandatory and cannot be applied “casually”. Section 45 is not soft now. It is contextual.

Tarsem Lal (2024): the summons-not-arrest carve-out

Of the four 2024 rulings, the one with the largest practitioner impact is arguably the smallest in profile. Tarsem Lal v. Directorate of Enforcement Jalandhar Zonal Office, 2024 INSC 434, decided in May 2024, did not address bail directly. It addressed what happens after the Special Court takes cognisance and issues a summons to an accused who has not been arrested. The doctrinal result has reshaped Section 45 strategy more than any of the headline cases.

The procedural facts: cognisance, summons, no arrest

The accused had not been arrested during the ED’s investigation. The agency filed its prosecution complaint. The Special Court took cognisance and issued process. The accused appeared on summons. The question was whether the twin conditions applied to a bail application in that posture.

The holding: twin conditions don’t apply where there is no arrest

The bench held that when an accused has not been arrested during the investigation and appears before the Special Court only in response to a summons issued after cognisance, the twin conditions under Section 45 do not apply at all. The bail application in that posture falls to be decided under the general framework of Section 480 / Section 483 BNSS, applying the ordinary triple-test analysis. The ED also loses the power to arrest the accused after cognisance has been taken; if the agency wants custody, it must apply to the Special Court.

The practitioner arbitrage and why the ED’s response is shifting

The second-order effect of Tarsem Lal has been a meaningful procedural arbitrage. In cases where the ED’s arrest grounds are weak, defence practitioners now actively seek the cognisance-and-summons route. If the prosecution complaint can be filed without prior arrest, the twin conditions never trigger, and the bail application clears under the ordinary BNSS framework. That’s a structural shift, not a tactical one. It also explains why the ED’s recent practice has been to arrest earlier and more often, sometimes against weaker grounds, because arrest is the operative event that locks in the twin conditions.

A question that practitioners raise is whether the ED can investigate further after cognisance, or whether the case is effectively closed once the prosecution complaint is filed. Tarsem Lal is clear: further investigation is permissible under Section 173(8) CrPC (now Section 193(9) BNSS), and any supplementary complaint can be filed. But the agency cannot use that ongoing investigation as a basis for fresh arrest after cognisance. That’s a procedural floor.

The pitfall to avoid is treating Tarsem Lal as a blanket bail entitlement in every cognisance-and-summons case. It isn’t. If the Special Court records that the accused is non-cooperative, evading process, or otherwise giving cause for custody, the court can still order custody, and the bail application then proceeds under the twin conditions because the operative event (custody) has occurred. Read the order, not just the headnote.

Kanhaiya Prasad (2025) and the pendulum back: bail not granted casually

By early 2025, lower courts had begun applying the post-2024 quartet with what the Supreme Court considered insufficient care. The Union of India v. Kanhaiya Prasad, 2025 INSC 210 judgment, decided in February 2025, was the correction.

The facts: HC bail set aside by the SC

A High Court had granted bail to an accused under the PMLA, citing the post-2024 quartet for the proposition that the twin conditions should be applied with sensitivity to the accused’s right to liberty. The Enforcement Directorate appealed. The Supreme Court set aside the bail order. The bench held that the High Court had failed to apply the twin conditions at all, and that the rulings in Manish Sisodia v. Directorate of Enforcement, 2024 INSC 595, Prem Prakash v. Union of India through Directorate of Enforcement, 2024 INSC 637, and V. Senthil Balaji v. Deputy Director, Directorate of Enforcement, 2024 INSC 739 do not displace Section 45; they contextualise it.

How Kanhaiya Prasad is being read with Sisodia and Prem Prakash

Read together, the post-2024 quartet and Kanhaiya Prasad give the bail court a structured analysis. First, the twin conditions remain mandatory; the court must apply them. Second, the prosecutor must be heard. Third, the court must record reasons for satisfaction on both prongs. Fourth, in doing so, the court can take into account delay, prosecutorial readiness, and the absence of arrest, but it cannot skip the twin-conditions analysis altogether. Skip it, and the bail order is vulnerable to ED appeal.

Reconciling the apparent contradiction

A common community question is whether Sisodia and Kanhaiya Prasad contradict each other. The short answer? No, but the framing matters. Sisodia is a delay-specific application of the twin conditions; Kanhaiya Prasad is a procedural reminder that the analysis must still happen. A defence counsel who cites Kanhaiya Prasad for the proposition that bail is impossible is misreading the case. So is a counsel who cites Sisodia for the proposition that the twin conditions are dispensable. The two cases together produce a contextual-application doctrine, and that doctrine is what trial courts in 2025 and 2026 are applying.

The pitfall here is citation hygiene. Kanhaiya Prasad is sometimes cited by the prosecution as if it overruled the post-2024 quartet. It didn’t. The judgment expressly preserves the speedy-trial line of cases; it disciplines the lower court’s application of them. Defence counsel should be ready to make exactly that distinction when the prosecutor raises it.

The first proviso: women, minors, sick and infirm, sums under Rs.1 crore

Tucked into the bottom of Section 45(1) is a proviso that has produced more recent litigation than the main provision. It opens four narrow categories where the twin conditions can be relaxed at the court’s discretion: women, persons under 16, the sick or infirm, and persons accused of money laundering involving a sum of less than Rs. 1 crore. The proviso reads in permissive language (“may be released on bail”), and that single word, “may”, has been the subject of two Supreme Court rulings in barely 12 months.

The four proviso categories and the “may be” question

The proviso operates as a permission, not as an automatic gateway. For an applicant who falls within one of the four categories, the court has discretion to grant bail without insisting on the twin conditions. The discretion is genuine; it is not a back-door mandatory relief. That framing matters because applications drafted as if the proviso were mandatory routinely fail.

The proviso is discretionary

In Saumya Chaurasia v. Directorate of Enforcement, 2023 INSC 1073, the Court held that the first proviso is not automatic. The fact that an applicant is a woman, or sick, or accused of a sub-Rs.-1-crore offence, does not by itself entitle her to bail. The court must still weigh the seriousness of the alleged conduct, the strength of the prosecution material, and the bail factors generally. The proviso opens a discretion. It does not pre-decide it.

Educated and politically active women not excluded

Less than a year later, the Court was asked whether the discretion under the first proviso could be denied to a woman accused who was politically active and educated. Some High Courts had begun denying the proviso’s benefit to such applicants on the reasoning that the proviso was meant for vulnerable women. In Kalvakuntla Kavitha v. Directorate of Enforcement, 2024 INSC 632, the Court rejected that reading. The proviso applies to women as a category; the High Court cannot exclude an applicant from the category on the basis of her social standing, education, or political activity. The proviso’s discretion still applies; the discretion can still go against the applicant. But the categorical exclusion approach is wrong.

The sick and infirm proviso in practice

The sick-and-infirm proviso has been the least-used limb of the four, partly because it requires evidentiary substantiation that defence counsel often fail to assemble. In practice, an application invoking this proviso needs medical board certification (preferably from a government hospital), a treatment plan, and a clear nexus between the medical condition and the inability to undergo continued custody. Generic medical records produced by the family of the accused are not enough; trial courts have been reading the proviso narrowly, and the Delhi High Court in 2021 set a relatively high threshold.

The Rs. 1 crore threshold: how it actually operates

The Rs. 1 crore proviso is read in conjunction with the prosecution’s allegations on quantum, not the accused’s own admission. If the ED alleges proceeds of crime exceeding Rs. 1 crore, the proviso is unavailable, even where the defence disputes the figure. Practitioners challenging quantum at the bail stage need contemporaneous documentary evidence (bank statements, transaction records, third-party confirmations) to displace the prosecution’s pleaded figure. Most applications attempting a quantum challenge fail because the defence brings only oral assertions.

The doctrinal arc of the women proviso, in particular, has compressed into 24 months. Here’s how it reads when laid side by side:

Phase	Ruling	What changed
Pre-Dec 2023	High Courts divided; many trial courts treating the proviso as broadly available	Inconsistent application; some near-mandatory readings, some highly restrictive
Dec 2023	Saumya Chaurasia v. Directorate of Enforcement (2023)	Proviso is discretionary; “may” means may, not “shall”
Aug 2024	Kalvakuntla Kavitha v. Directorate of Enforcement (2024)	High Courts cannot categorically exclude educated or politically active women from the proviso’s reach
Late 2025 onwards	Lower-court applications	Discretionary, with stricter scrutiny of the underlying allegations; politically exposed applicants no longer auto-excluded

The pitfall in litigating the proviso is treating it as automatic relief. Post-Saumya Chaurasia, this misreading has cost defence applicants real bail orders, and it remains the single most common error in proviso-based applications today.

Section 45 PMLA vs Section 37 NDPS vs Section 43D UAPA: bail bars compared

PMLA is one of three major Indian statutes that layer a special bail bar on top of the BNSS / CrPC framework. The other two are the Narcotic Drugs and Psychotropic Substances Act, 1985 (Section 37) and the Unlawful Activities (Prevention) Act, 1967 (Section 43D(5)). Practitioners regularly mix up the three, and the prosecutor uses that confusion. A careful comparison makes both arguments and counter-arguments cleaner.

Trigger and scope: which offences attract which bar

Section 45 PMLA attracts on any offence under the Act after the 2018 amendment. Section 37 NDPS attracts on commercial-quantity offences only; intermediate-quantity and small-quantity offences are governed by the ordinary BNSS framework. Section 43D(5) UAPA attracts on offences under Chapters IV and VI of the Act (terrorism, conspiracy, unlawful association), not on every offence under the UAPA. The triggering offence matters; an application under the wrong framework gets dismissed for misconception.

The court’s satisfaction test: how the language differs

The cleanest way to see the three bars side by side is in a comparison table, because the textual differences are small but consequential.

Feature	Section 45 PMLA, 2002	Section 37 NDPS Act, 1985	Section 43D(5) UAPA, 1967
Trigger for the bail bar	Any offence under PMLA, post-2018 amendment	Commercial-quantity narcotic offences	Chapter IV and VI offences (terrorism, conspiracy)
Public Prosecutor’s role	Mandatory opportunity to oppose	Mandatory opportunity to oppose	No equivalent express requirement
Court’s satisfaction test	Reasonable grounds to believe accused not guilty plus unlikely to commit offence on bail	Reasonable grounds to believe accused not guilty plus unlikely to commit offence on bail	Court of opinion that accusation is prima facie not true
Provisos (women, minor, sick, threshold)	Yes: women, persons under 16, sick or infirm, sums under Rs.1 crore	Persons under 18 and women only (limited)	No corresponding provisos
Standard at bail stage	Twin conditions cumulative; both must be met	Twin conditions cumulative; both must be met	Single prong: prima facie truth of accusation
Anticipatory bail (Section 438 CrPC / 482 BNSS)	Section 45 conditions apply	Section 37 conditions apply	Section 43D applies; anticipatory bail effectively unavailable
Speedy-trial relief	Read in by Sisodia, Senthil Balaji, Prem Prakash (2024)	Recognised in scattered HC rulings	Recognised in K.A. Najeeb (2021): constitutional courts can grant bail despite Section 43D

The textual difference between PMLA and NDPS is small; the practical difference between either and UAPA is large. Under PMLA and NDPS, the bail court must affirmatively find that the accused is not likely to commit an offence on bail. Under UAPA, the court doesn’t apply that prong; it asks only whether the accusation is prima facie true. That makes UAPA bail mechanically harder in the abstract but easier in cases where the prosecution material is thin, because there is no second prong the prosecutor can hide behind.

Provisos and proviso architecture

The proviso architecture differs across all three. PMLA has the four-category first proviso. NDPS has a similar but narrower proviso. UAPA has none. A defence counsel who reflexively imports proviso arguments from PMLA into UAPA litigation is making a textual error; the UAPA bench will dismiss the argument on threshold.

Speedy-trial doctrine across all three

What experienced practitioners know is that the speedy-trial doctrine has had different lives in each statute. UAPA had its breakthrough in the 2021 ruling that constitutional courts can grant bail despite Section 43D where the trial is delayed. PMLA had its breakthrough in 2024 through the post-2024 quartet. NDPS has not yet had a clear Supreme Court ruling on speedy-trial-based bail; the High Courts are divided. A second-order effect is that PMLA bail strategy is currently more developed than NDPS bail strategy, and counsel who practise in both should not assume the PMLA toolkit transfers cleanly.

A community question that arises here is how PMLA compares with MCOCA and TADA bail provisions. MCOCA Section 21(4) is similar in structure to Section 45 PMLA but slightly narrower; TADA’s Section 20(8) was harsher and is no longer operative as TADA has lapsed. None of these statutes import the post-2024 quartet automatically. Practitioners arguing analogies need to be careful.

The pitfall in this comparison is the assumption that “twin conditions” mean the same thing across statutes. They don’t. The language is similar; the doctrinal weight differs. Read the statute, then read the case law on that statute, before drawing analogies.

Section 45 PMLA vs Section 37 NDPS vs Section 43D UAPA

Bail bars compared across three special statutes

Feature	Section 45 PMLA, 2002	Section 37 NDPS, 1985	Section 43D(5) UAPA, 1967
Trigger for the bail bar	Any offence under PMLA (post-2018 amendment)	Commercial-quantity narcotic offences	Chapter IV and VI offences (terrorism, conspiracy)
Public Prosecutor’s role	Mandatory opportunity to oppose	Mandatory opportunity to oppose	No equivalent express requirement
Court’s satisfaction test	Reasonable grounds to believe accused not guilty + unlikely to commit offence on bail	Reasonable grounds to believe accused not guilty + unlikely to commit offence on bail	Court of opinion that accusation is prima facie not true (no “not likely” prong)
Provisos (women / minor / sick / threshold)	Yes – women, persons under 16, sick or infirm, sums under Rs.1 crore	Persons under 18 and women only (limited)	No corresponding provisos
Standard at bail stage	Twin conditions cumulative; both must be met	Twin conditions cumulative; both must be met	Single prong: prima facie truth of accusation
Anticipatory bail (S.438 CrPC / S.482 BNSS)	Section 45 conditions apply (Gopal Reddy, 2023)	Section 37 conditions apply at the pre-arrest stage	Section 43D applies; anticipatory bail effectively unavailable in chargesheeted Chapter IV/VI cases
Speedy-trial relief	Read in by Sisodia, Senthil Balaji, Prem Prakash (2024)	Recognised in scattered HC rulings; no SC equivalent	K.A. Najeeb (2021): constitutional courts can grant bail despite Section 43D

Section 45 PMLA, 2002

Trigger

Any offence under PMLA (post-2018 amendment)

Prosecutor’s role

Mandatory opportunity to oppose

Satisfaction test

Reasonable grounds to believe accused not guilty + unlikely to commit offence on bail

Provisos

Women, persons under 16, sick or infirm, sums under Rs.1 crore

Standard

Twin conditions cumulative

Anticipatory bail

Section 45 conditions apply (Gopal Reddy, 2023)

Speedy-trial relief

Read in by Sisodia, Senthil Balaji, Prem Prakash (2024)

Section 37 NDPS Act, 1985

Trigger

Commercial-quantity narcotic offences

Prosecutor’s role

Mandatory opportunity to oppose

Satisfaction test

Reasonable grounds to believe accused not guilty + unlikely to commit offence on bail

Provisos

Persons under 18 and women only (limited)

Standard

Twin conditions cumulative

Anticipatory bail

Section 37 conditions apply at the pre-arrest stage

Speedy-trial relief

Scattered HC rulings; no SC equivalent

Section 43D(5) UAPA, 1967

Trigger

Chapter IV and VI offences (terrorism, conspiracy)

Prosecutor’s role

No equivalent express requirement

Satisfaction test

Court of opinion that accusation is prima facie not true (no “not likely” prong)

Provisos

None

Standard

Single prong: prima facie truth of accusation

Anticipatory bail

Effectively unavailable in chargesheeted Chapter IV/VI cases

Speedy-trial relief

K.A. Najeeb (2021): constitutional courts can grant bail despite Section 43D

Compare carefully: the bail bars differ in trigger, test, and proviso architecture. The language is similar; the doctrinal weight differs.

Anticipatory bail under Section 438 CrPC / Section 482 BNSS: does Section 45 apply?

A perennial question for those facing the prospect of an ED arrest is whether anticipatory bail is available, and if so, whether the twin conditions of Section 45 apply at the pre-arrest stage. The doctrinal answer has crystallised through one Supreme Court ruling, but defence practitioners continue to argue around its edges.

The position before Gopal Reddy

Before 2023, the High Courts were divided. Some took the view that Section 45’s twin conditions applied only to post-arrest bail under Section 437 / 439 CrPC, and that anticipatory bail under Section 438 of the Code of Criminal Procedure, 1973 (now Section 482 of the Bharatiya Nagarik Suraksha Sanhita, 2023) fell under the general framework without Section 45 layered on top. Other High Courts read the Act’s purpose as requiring the twin conditions to apply at every bail-grant stage, including pre-arrest. Defence counsel exploited the inconsistency for forum-friendly applications.

The closure of the backdoor

In Directorate of Enforcement v. M. Gopal Reddy, 2023 LiveLaw (SC) 138, the Supreme Court resolved the split. The two-judge bench held that the twin conditions under Section 45 apply to anticipatory bail applications in PMLA cases. The pre-arrest avenue was, in effect, closed; the same threshold that applies after arrest applies before. The reasoning was that if Section 45 were inapplicable to anticipatory bail, the bar would be illusory because an accused could simply move ahead of an arrest and obtain pre-arrest relief on lower thresholds.

What survives: pre-arrest protection on narrow grounds

What remains available is pre-arrest protection on the same twin-conditions analysis. That means the application has to make a prima facie case on both prongs at the pre-arrest stage, often without the benefit of seeing the ECIR or knowing the precise allegations. That’s a steep ask. In practice, anticipatory bail in PMLA cases is granted sparingly, and primarily where the apprehended arrest is patently arbitrary, the predicate offence has been quashed, or the applicant has cooperated extensively and can show that detention is not reasonably necessary.

In practice, what experienced practitioners do is structure the pre-arrest application as a hybrid: a Section 438 CrPC / Section 482 BNSS prayer, layered on a Section 45 twin-conditions argument, supported by an affidavit demonstrating cooperation and absence of flight risk. The pitfall is treating Gopal Reddy as making anticipatory bail unavailable. It doesn’t. It makes it harder, but not impossible.

Default bail under Section 187(2) BNSS / Section 167(2) CrPC: still available in PMLA?

Default bail (sometimes called statutory bail or compulsive bail) is the indefeasible right to bail that accrues to an accused when the investigation is not completed within the period specified by statute. The general framework is now in Section 187 of the Bharatiya Nagarik Suraksha Sanhita, 2023 (which corresponds to the older Section 167(2) CrPC). The question for PMLA is whether this right operates independently of, or is subordinated to, the twin conditions.

The 60 / 90-day rule and how it applies to PMLA

Under Section 187 BNSS, an accused has the right to default bail if the investigating agency does not file the chargesheet within 60 days for offences with a maximum punishment of less than 10 years, and within 90 days for more serious offences. PMLA offences carry a maximum imprisonment of seven years (10 years where the predicate is a Narcotic Drugs and Psychotropic Substances Act offence). For most PMLA offences, the 60-day window applies; for the narcotics-predicate cases, 90 days.

Crucially, default bail is not subject to the twin conditions. Once the right has accrued, and the accused has applied (the right has to be claimed; it doesn’t operate automatically), the court must release the accused on bail. Section 45 PMLA does not override Section 187 BNSS / Section 167(2) CrPC. The Supreme Court has been firm on this point both before and after Vijay Madanlal Choudhary.

ED chargesheet timelines and judicial discipline post-Vijay Madanlal

Post-Vijay Madanlal, courts have tightened the discipline on default bail timing. The ED must file its prosecution complaint within the statutory window, and any cognisance taken after the window without an application by the accused for default bail extinguishes the right (the right has to be exercised the moment it accrues). In practice, this means defence counsel must watch the calendar precisely. Missing the default-bail window because the accused was waiting for a substantive bail outcome is a real and recurring practitioner error.

The pitfall here is treating default bail as a fallback. It isn’t. It’s a primary tool when the ED is running tight on chargesheet timing. And in complex PMLA investigations, with cross-jurisdictional sums and multiple co-accused, the ED frequently runs tight.

Article 21, speedy trial, and the 2026 wherewithal test

The constitutional spine of post-2024 PMLA bail jurisprudence is Article 21. Every softening of the twin conditions has been justified on the right to a speedy trial and the right to liberty. In January 2026, the Supreme Court added a new variable that practitioners are still working out: prosecutorial wherewithal.

How the speedy-trial doctrine is read into Section 45

The 2024 quartet established a framework: where the trial has not begun within a reasonable period (Sisodia’s 17 months was the high-water mark, but later orders have applied it at shorter durations), and where there is no realistic prospect of imminent commencement (Senthil Balaji), constitutional courts must intervene under Article 21. The intervention happens through Section 45 itself, not as an override. The bail court treats Article 21 as a reading-in principle: when the twin conditions are applied in light of the right to liberty, prolonged custody without trial progress weighs heavily on the second prong.

The January 2026 wherewithal ruling: prosecution capacity as a bail variable

In Arvind Dham v. Directorate of Enforcement, 2026 INSC 12, decided on 6 January 2026, a two-judge bench of the Supreme Court granted bail to a PMLA accused on the basis that the prosecution lacked the wherewithal to complete the trial within a reasonable time. The bench introduced a new analytical variable: instead of asking only how long the accused has been in custody, the court also asked whether the prosecution had the institutional and procedural capacity to bring the trial to a conclusion. If the answer is no, the twin conditions cannot operate to prolong detention indefinitely.

This is a sharper test than Sisodia. Sisodia asked retrospective questions; the wherewithal test adds a prospective one. And it shifts at least part of the burden onto the prosecution to show readiness, not just willingness, to proceed.

What 2027 to 2028 may bring: legislative pressure and the Chhattisgarh HC nudge

Looking forward, two signals are worth tracking. First, the March 2025 observation by the Chhattisgarh High Court that Parliament should consider expressly carving out case-by-case discretion under Section 45. That observation echoes academic commentary from the late 2024 NUALS Law Journal piece and may, over time, produce legislative momentum. Practitioners expect that any amendment, if it comes, will travel through a Standing Committee referral first, which would itself be a signal worth watching.

Second, the wherewithal test is likely to expand. Early signals suggest High Courts in metros are beginning to ask the ED to provide a chargesheet-readiness timeline as part of the bail-hearing record. That’s a procedural shift the post-2024 quartet implies but does not formally require. Practitioners drafting bail applications in 2026 and beyond should anticipate the question and pre-empt it through a tabulated custody-and-trial-progress schedule.

A community question that surfaces here is what the maximum period of custody is before the court must release the accused. There is no fixed maximum under PMLA. The constitutional framework asks for a reasonable trial timeline; “reasonable” is contextual. Practitioners who promise their clients a numerical guarantee misrepresent the doctrine. The accurate framing is that the longer the custody and the slower the trial, the stronger the speedy-trial argument, with 12 to 18 months emerging as the practical inflection point in most fora.

ECIR, Section 50 statements, and predicate offence interaction

Section 45 doesn’t operate in a vacuum. Its application turns on the ECIR (Enforcement Case Information Report), the Section 50 statements of the accused and witnesses, and the status of the underlying predicate offence. Defence practitioners who treat these as separate workstreams miss the integration. The strongest right to bail in money-laundering cases applications work all three.

What an ECIR is and whether the accused has a right to see it

The ECIR is the ED’s internal document recording the registration of a money-laundering investigation. Unlike an FIR, it is not a public document, and the Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929 court held that the accused has no fundamental right to a copy. The accused is, however, entitled to be informed of the grounds of arrest under Section 19 PMLA, and post-Pankaj Bansal (2023) and Prabir Purkayastha (2024) those grounds must be furnished in writing.

The practical consequence is that defence counsel preparing for a bail hearing usually do not have the full ECIR, only the grounds of arrest, the prosecution complaint if filed, and any annexures supplied by the prosecution at the hearing. Application drafting has to work around this informational asymmetry.

Section 50 statements at the bail stage: admissibility and use

Statements recorded under Section 50 PMLA, by persons summoned by the ED, are admissible in evidence under the Act. At the bail stage, the prosecution often relies on them to establish the not-likely-to-commit prong. Defence counsel have a difficult choice: contesting the admissibility of these statements at bail stage is rarely successful, but they can be neutralised by pointing out internal inconsistencies, retractions, or the absence of corroboration.

The Prem Prakash bench specifically commented on the use of Section 50 statements, emphasising that the prosecution cannot bootstrap the entire case on the accused’s own statement when challenged. A second-order effect is that the cooperation calculus has flipped. Pre-2024, cooperation was sometimes used by the ED as evidence of guilt-consciousness. Post-Prem Prakash, courts increasingly treat cooperation as material to the not-likely-to-commit prong, which means the defence advice on whether to attend Section 50 summons has been changing.

When the predicate offence is quashed: does PMLA survive?

This is one of the most-litigated and least-blogged questions in current PMLA practice. The general rule, derived from Vijay Madanlal, is that the PMLA case requires a scheduled offence at its base. If the predicate offence is quashed by a competent court (whether through a Section 482 BNSS quashing petition, discharge, or acquittal), the PMLA case is fundamentally weakened, though courts have been careful not to say it automatically collapses. There is also a deep companion piece on how predicate offences under the PMLA are identified in practice that is worth reading before any predicate-strategy argument.

What practitioners increasingly do is run parallel quashing petitions for the predicate alongside the bail application. If the quashing is granted, the bail case strengthens dramatically. If it is even seriously argued, the bail court has discretion to treat the predicate’s vulnerability as a bail factor. This parallel-quashing strategy is under-discussed in general explainers; in current practice, it is one of the most effective indirect routes to PMLA bail. The Certificate Course in Banking and Finance Laws covers the underlying white-collar predicate frameworks (such as SARFAESI and the Companies Act provisions) that frequently underlie PMLA cases.

Pavana Dibbur and the named-in-FIR question

A practitioner question that frequently arises is whether a person can be arrested under PMLA if they have not been named in the FIR for the scheduled offence. The Supreme Court has clarified that the PMLA case does not require the accused to be named in the predicate FIR. What matters is whether the accused has dealt with proceeds of crime as defined under Section 2(1)(u) PMLA. That said, an accused not named in the predicate FIR has a structurally stronger argument that the foundational facts are missing, and the Prem Prakash framework reinforces this.

The pitfall here is conflating the predicate FIR with the ECIR. The ECIR is the ED’s internal record; the FIR is the predicate offence record kept by the regular police agency. They are separate, and a quashing of one does not automatically operate on the other.

Procedural workflow: from ED arrest to first bail hearing

Most accused persons and their families want a procedural map: what happens when, who decides what, and how soon can a bail application even be filed. The PMLA workflow has its own quirks compared with ordinary criminal procedure, and the first 48 to 96 hours after arrest are particularly consequential.

Arrest under Section 19 PMLA and the grounds-of-arrest requirement

ED arrests are governed by Section 19 PMLA. The investigating officer must record reasons for arrest and inform the arrested person of the grounds. Following the 2023 Pankaj Bansal ruling and the 2024 Prabir Purkayastha ruling, those grounds must be furnished in writing to the arrested person and their counsel, not merely read out. A failure to comply with this requirement is by itself a ground for the High Court to set aside the arrest as illegal, which in turn collapses the bail context altogether.

Production before the Special Court and the remand cycle

The arrested person must be produced before the jurisdictional Special Court (or the nearest Magistrate, depending on geography) within 24 hours. The ED typically applies for ED custody for an initial period (commonly 10 to 14 days), after which the accused moves to judicial custody. Bail applications can be filed at any point after arrest, but in practice, the first effective window is after the ED custody phase ends and the accused is in judicial custody, because that’s when the defence has access to the grounds-of-arrest material and can examine the remand record.

ED chargesheet timing and the first effective bail window

The ED has 60 days (or 90 days, in narcotics-predicate cases) to file a prosecution complaint. Defence counsel typically prepare two parallel tracks: an early substantive bail application based on the grounds of arrest and the absence of foundational facts, and a fallback default-bail application timed to the expiry of the statutory window. A community question that families raise is how soon they can apply for bail. The procedural answer is that an application can be filed immediately after arrest; the strategic answer is that the first hearing has the strongest chance after the ED custody phase and before the prosecution complaint is filed.

Jurisdiction: Special Court, Sessions Court, or High Court?

PMLA bail applications must be filed before the designated Special Court (the Court of Sessions designated under Section 43 PMLA) in the first instance. Going directly to the High Court without first approaching the Special Court is procedurally improper in most circuits, though some High Courts will entertain a writ-route bail in exceptional cases (illegal arrest, gross delay, custodial torture allegations). The default workflow is Special Court first, High Court on appeal or revision.

What experienced practitioners know is that the role of the Public Prosecutor at the Section 45 hearing is procedural and substantive. Procedurally, the prosecutor must be given an opportunity to oppose; the failure to give that opportunity is by itself a ground for the order to be set aside on appeal. Substantively, the prosecutor’s submissions frame the not-likely-to-commit analysis. Defence counsel who under-prepare for the prosecutor’s specific submissions, rather than for the general prosecution case, regularly underperform at the bail stage.

Practitioner’s checklist for drafting a Section 45 PMLA bail application

This is where the article shifts from doctrinal explainer to practical drafting guide. The checklist below isn’t a substitute for case-specific drafting; it’s a working sheet to ensure no major piece is missed before the bail application is filed. The broader mechanics of drafting a bail application under the BNSS are covered separately in our forthcoming companion piece on BNSS bail drafting; here, the focus is the PMLA-specific layer.

Pre-filing diligence

Before the application is even drafted, four pieces of paperwork should be on counsel’s desk. First, the ECIR copy if obtainable (a formal request should be on record even if the ED has refused). Second, the predicate offence FIR and any chargesheet, plus the status of any quashing or discharge proceeding. Third, the arrest memo and the written grounds of arrest under Section 19 PMLA, scrutinised for Pankaj Bansal and Prabir Purkayastha compliance. Fourth, the full remand record (every order from production through the current date). Add to this any bail orders already passed for co-accused, which are essential for the parity argument.

A practitioner pitfall at this stage is over-relying on the family for documents. Family-supplied records are often incomplete, sometimes selectively curated, and almost always missing the prosecution-side annexures. Direct certified copies from the court record are worth the procedural delay.

Framing the application: title, prayer, grounds

The application should be titled “Application under Section 45 of the Prevention of Money Laundering Act, 2002 read with Section 483 of the Bharatiya Nagarik Suraksha Sanhita, 2023” (for an application before the Sessions or High Court). The prayer should be specific (release on regular bail on furnishing surety, with conditions to be imposed by the Court). The grounds should be structured: first the prima-facie-absence-of-foundational-facts argument, then the twin-conditions analysis prong by prong, then the speedy-trial / Article 21 argument if applicable, then the proviso analysis if applicable, then the parity argument if applicable, then a list of proposed bail conditions the applicant is willing to undertake.

Arguing the twin conditions: prong-by-prong tactics

On the reasonable-grounds prong, the strongest framing is that the prosecution material, taken at face value, does not establish a prima facie nexus between the applicant and proceeds of crime. The Prem Prakash burden-shifting argument anchors this. On the not-likely-to-commit prong, the applicant should pre-emptively offer the conditions menu: passport surrender, residence in jurisdiction, daily or weekly reporting at the IO’s office, no contact with witnesses or co-accused other than through counsel, no foreign travel. Pre-emptive condition-offering shifts the conversation from whether bail is appropriate to what conditions are appropriate, and that’s a much better conversation for the defence.

Co-accused parity, cooperation, and Section 50 neutralisation

Where co-accused have been granted bail, the parity argument is mandatory; not raising it concedes a point that is hard to recover on appeal. Cooperation should be documented (Section 50 summons attended, statements furnished, documents produced). The Prem Prakash framework treats cooperation as supportive of the not-likely-to-commit prong, so contemporaneous evidence of cooperation is now actively helpful, where pre-2024 it was a more ambivalent factor. Section 50 statements relied on by the prosecution should be neutralised through internal inconsistencies, retractions, or absence of corroboration; outright admissibility challenges at bail stage usually fail.

NRIs, foreign nationals, and passport-surrender architecture

NRI and foreign-national applicants have additional structural challenges, particularly on the flight-risk dimension. The passport surrender is necessary but rarely sufficient. The practitioner approach is to layer surrender with a fresh address in jurisdiction (frequently a relative’s residence), a willingness to deposit a bond proportional to the offence, and an undertaking on reporting frequency higher than the standard. For foreign nationals, an additional consular intimation clause is increasingly accepted by metro Special Courts as a substitute for some of the more onerous conditions.

A second-order effect of the post-2024 framework is that the bail application now does double duty. It is the bail application, and it is also the first substantive engagement with the prosecution’s case, which the Special Court will read at trial. Concessions made or inferences invited at the bail stage carry over. Defence counsel who treat the bail application as a self-contained document, with no thought to trial consequences, regularly lose ground they cannot easily recover.

This is exactly the kind of bail-application drafting craft covered in depth in LawSikho’s Diploma in Criminal Litigation and Trial Advocacy: section-wise drafting frameworks, twin-conditions argumentation, and live cross-examination practice on real-style PMLA briefs.

Practitioner’s Section 45 PMLA bail application checklist

Four stages, eighteen actionable steps

1 Pre-filing diligence

1. ECIR copy applied for and on record (or denial grounds noted).
2. Predicate FIR / chargesheet annexed; quashing status checked.
3. Arrest memo + Section 19 grounds tested for Pankaj Bansal and Prabir Purkayastha compliance.
4. Custody, remand orders, and judicial-custody dates tabulated.
5. Co-accused bail orders collected for the parity argument.

2 Framing the bail application

1. Title: “Application under Section 45 PMLA r/w Section 483 BNSS / Section 439 CrPC”.
2. Identify any qualifying proviso: woman / under-16 / sick or infirm / sum under Rs.1 crore.
3. Plead Article 21 speedy-trial ground if pre-trial custody exceeds 12 months.
4. Cite Sisodia, Prem Prakash, Senthil Balaji (2024) for delay doctrine; distinguish Kanhaiya Prasad (2025).
5. If summons-only after cognisance, invoke Tarsem Lal (2024) – twin conditions inapplicable.

3 Arguing the twin conditions

1. Reasonable-grounds prong: show absence of mens rea or proceeds-of-crime nexus (Prem Prakash burden).
2. Not-likely prong: pre-offer passport surrender, reporting schedule, no-contact, residence-in-jurisdiction.
3. Neutralise Section 50 statements via inconsistencies or retractions (per Prem Prakash, Senthil Balaji).
4. Avoid factual concessions the Special Court may treat as binding at trial.

4 Post-grant compliance

1. Furnish surety; surrender passport; report at IO’s office on schedule.
2. Maintain residence in jurisdiction; intimate any travel.
3. No contact with witnesses or co-accused except through counsel.
4. Track ED’s appeal window – Section 45(2) read with BNSS bail-cancellation law.

Use this as a working sheet, not a substitute for case-specific drafting.

Common mistakes in PMLA Section 45 bail litigation

Across hundreds of bail orders read between 2022 and 2026, five errors recur often enough to be worth flagging separately. Each is avoidable. Each costs defence applicants meaningful bail orders.

Treating the proviso as automatic relief

Post-Saumya Chaurasia v. Directorate of Enforcement, 2023 INSC 1073, the first proviso is discretionary. Applications drafted as if the woman, sick-infirm, or sub-Rs.-1-crore proviso were a guaranteed gateway routinely fail. The correct framing is: the applicant qualifies under the proviso, and the court has discretion to relax the twin conditions in their case, and the discretion should be exercised because (specific factors). Drop the “should be exercised because” portion and the application reads as entitlement-based.

Relying on cooperation alone without prong-by-prong framing

Cooperation with the investigation is helpful but not sufficient. Applications that lead with cooperation, and only later get to the twin-conditions analysis, signal to the bench that counsel is hoping cooperation will substitute for the statutory test. It won’t. The structure should be reversed: prong-by-prong analysis first, with cooperation woven into the not-likely-to-commit prong.

Citing the wrong case for the wrong proposition

Union of India v. Kanhaiya Prasad, 2025 INSC 210 is sometimes cited by the defence on the assumption that any 2025 Supreme Court ruling on PMLA must be helpful. It is, in fact, a prosecution-friendly case. Citing it carelessly hands the prosecutor a ready response. Conversely, Manish Sisodia v. Directorate of Enforcement, 2024 INSC 595 is sometimes cited for the proposition that the twin conditions are dispensable, which is a misreading and invites the prosecutor to point out that Sisodia preserved the twin conditions, applied them, and granted bail on the speedy-trial reading.

Ignoring the BNSS renumbering

Most practitioner templates still read “Section 437 / 439 CrPC”. With the BNSS in force since 1 July 2024, the correct citations are Section 480 BNSS and Section 483 BNSS. A small detail. But it dates the application, and judges notice. The CrPC parenthetical is fine; leading with CrPC vocabulary signals that the file hasn’t been updated since the BNSS rollout. A second-order effect is that as the BNSS jurisprudence develops (rules on default bail, on transitional provisions for cases pending on 30 June 2024), counsel who still draft in CrPC vocabulary will miss the cleaner statutory hooks.

Failing to pre-empt the prosecutor’s specific arguments

The single most common error across bail applications is treating the prosecutor as a generic opponent. Defence counsel who fail to anticipate the prosecutor’s specific paragraphs (which Section 50 statement will be quoted, which witness is allegedly at risk of intimidation, which financial transaction is being relied on for the foundational facts) under-prepare the application. The fix is procedural: collect the prosecutor’s prior submissions in the same case (remand applications, opposition to prior bail) and pre-empt them paragraph by paragraph.

Recent Supreme Court and High Court trends: what 2025-2026 has changed

The case law is moving fast. A bail application drafted on early 2024 precedent reads differently from one drafted on early 2026 precedent, even where the facts are similar. Two clusters of recent rulings are worth tracking.

The 2025 cohort: Kanhaiya Prasad, Arvind Dham, and the wherewithal test

The 2025 Supreme Court output on PMLA bail has been mixed in direction but consistent in one signal: the Court wants the twin conditions applied rigorously but contextually. The Kanhaiya Prasad ruling reinforced rigour. The Arvind Dham line of rulings (and the parallel “wherewithal” judgment) reinforced context. Together they signal a doctrinal equilibrium: neither pure strictness nor pure liberty, but reasoned application with full record.

Trial courts that read this cohort carefully are producing more granular orders. Some grant bail with intensive conditions packages; others deny but record reasons that signal what would tip the next application towards grant. Defence counsel benefit from reading these orders for the second application, not just the case law.

The 2026 cohort: Delhi HC denials in cyber-fraud PMLA cases

Early 2026 has seen the Delhi High Court denying anticipatory bail in cyber-fraud PMLA cases on strict twin-conditions reasoning. The pattern is fact-specific (typically large alleged sums, multiple jurisdictions, evidence of ongoing fraud during the investigation) but doctrinally instructive. The High Court has emphasised that the post-2024 quartet does not soften the bar in cases with active conduct allegations. Counsel arguing speedy-trial relief in cyber-fraud cases need to address this distinction at the front of the application.

Pattern: contextual application, not formulaic application

The cleanest synthesis of 2025 and 2026 is that the bail court is doing more work than it used to. Pre-2022 orders were short; the twin conditions did most of the analytical work. Post-2024 orders are longer, more reasoned, and more record-dependent. That has implications for application drafting: provide the bench with the analytical scaffolding it now wants. A two-page application that worked in 2020 will lose in 2026.

Bail conditions, violations, and ED appeals

Once bail is granted, the operative work shifts to compliance and risk management. The conditions imposed in PMLA bail orders are heavier than ordinary bail, and the consequences of violation are correspondingly serious.

Typical conditions imposed

Standard PMLA bail conditions include passport surrender (often deposited with the Court), reporting at the investigating officer’s office at specified intervals (weekly is common in early-stage bail; monthly in post-chargesheet bail), residence in jurisdiction, no contact with witnesses or co-accused except through counsel, no foreign travel without prior court permission, and intimation of any change in residence or contact details. Some Special Courts add surety amounts proportional to the alleged proceeds of crime; others require fresh affidavits at periodic intervals confirming compliance.

Consequences of violation

Violation of bail conditions is independently actionable. The ED can apply for cancellation of bail under Section 45(2) read with the BNSS bail-cancellation framework. Cancellation does not require fresh charges; it requires only proof that conditions were violated and that the violation undermines the bail premise. Courts have been particularly strict on witness contact and on foreign travel without permission. A pitfall here is the assumption that minor procedural breaches will be excused. They sometimes are; they sometimes aren’t. The risk-adjusted answer is to err on the side of over-compliance.

ED’s appeal route against bail orders

The ED can appeal a bail order under the appellate framework applicable to the level of court that passed it. Special Court bail orders are appealable to the High Court; High Court bail orders to the Supreme Court. The window is short; counsel should expect an appeal within 30 to 60 days for any politically or commercially significant bail grant. Defence counsel should prepare a counter-affidavit pack at the same time as the bail application, anticipating the appeal rather than reacting to it.

Practitioners who master PMLA bail jurisprudence command premium briefs in white-collar criminal defence. LawSikho’s Diploma in Criminal Litigation and Trial Advocacy has been built for exactly this expertise, from first-year associates running their first bail brief to mid-career advocates building a specialist defence practice.

Frequently asked questions on PMLA Section 45 bail

What are the twin conditions for bail under Section 45 of the PMLA?

The twin conditions are cumulative: first, the public prosecutor must be given an opportunity to oppose the bail application; second, the court must be satisfied that there are reasonable grounds for believing the accused is not guilty of the offence and is not likely to commit any offence while on bail. Both prongs must be met before bail can be granted. The standard is prima facie, not beyond reasonable doubt.

Is money laundering under the PMLA a bailable or non-bailable offence?

Offences under the PMLA are cognisable and non-bailable. That classification means there is no automatic right to bail; the accused must apply, and the court has discretion. Section 45 then layers additional twin conditions on top of the general BNSS bail framework, making PMLA bail materially harder than ordinary non-bailable bail. The label is the start of the analysis, not the end.

Why is it so difficult to get bail under the PMLA?

PMLA bail is difficult for three structural reasons. First, the cumulative twin conditions under Section 45 require the court to be satisfied on a not-guilty prong at the bail stage. Second, the prosecution can rely on Section 50 statements at the bail hearing, statements that are admissible under the Act. Third, Section 24 PMLA imports a reverse presumption on proceeds of crime, shifting some of the evidentiary burden onto the accused even at preliminary stages.

Are the twin conditions in Section 45 still constitutional after Nikesh Tarachand Shah?

Yes. Although Nikesh Tarachand Shah v. Union of India, (2018) 11 SCC 1 struck down the original Section 45(1), Parliament amended the provision through the Finance Act, 2018. The 2018-amended Section 45 was upheld as constitutional by a three-judge bench of the Supreme Court in Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929. The twin conditions are now firmly back, though subsequent rulings have softened their application in delay-based and summons-based contexts.

What did the Supreme Court hold in Vijay Madanlal Choudhary (2022) on Section 45?

A three-judge bench in Vijay Madanlal Choudhary upheld the 2018-amended Section 45 as constitutional. The Court characterised money laundering as a heinous crime with grave socio-economic consequences and held that the twin conditions, applied in light of the gravity of the offence, do not violate Articles 14 or 21. The judgment also upheld other PMLA provisions, including Section 19 arrest powers and Section 50 statements.

Can a woman accused under the PMLA get bail automatically under the first proviso?

No. The first proviso to Section 45(1) is discretionary, not mandatory. The Supreme Court in Saumya Chaurasia v. Directorate of Enforcement, 2023 INSC 1073 clarified that the word “may” in the proviso means may, not shall. The court can relax the twin conditions for a woman applicant, but it is not required to. The discretion must be exercised on the case-specific facts, and the proviso is not available as automatic relief.

Can a sick or infirm person get bail under Section 45 PMLA?

The sick or infirm proviso under Section 45(1) is available, but it requires substantial medical evidence. Generic medical records from a family physician are rarely sufficient. The application typically needs a government hospital medical board certification, a documented treatment plan, and a clear nexus between the medical condition and the inability to undergo continued custody. Even then, the proviso is discretionary, not automatic.

What is the Rs. 1 crore threshold proviso under Section 45 PMLA?

The first proviso permits the court to consider bail without applying the twin conditions where the alleged proceeds of crime involve a sum of less than Rs. 1 crore. The threshold is read against the prosecution’s pleaded figure, not the accused’s denial. Practitioners challenging quantum at the bail stage need contemporaneous documentary evidence to displace the pleaded figure. The proviso remains discretionary at all stages.

Can default bail under Section 187(2) BNSS / Section 167(2) CrPC apply in PMLA cases?

Yes. Default bail is available in PMLA cases when the ED fails to file the prosecution complaint within 60 days (or 90 days for narcotics-predicate cases). Default bail is not subject to the twin conditions; once the right has accrued and is exercised, the court must release the accused. The right must be claimed at the right time; delay in claiming can extinguish it, so timing discipline matters.

Are the twin conditions attracted when an accused appears on summons rather than on arrest?

No. In Tarsem Lal v. Directorate of Enforcement Jalandhar Zonal Office, 2024 INSC 434, the Supreme Court held that the twin conditions under Section 45 do not apply when the accused has not been arrested during the investigation and appears before the Special Court only in response to a summons issued after cognisance. The bail application in that posture proceeds under the ordinary BNSS bail framework. The agency also loses the power to arrest the accused after cognisance.

Does Section 45 PMLA apply to anticipatory bail under Section 438 CrPC / Section 482 BNSS?

Yes. The Supreme Court has held that the twin conditions of Section 45 apply equally to anticipatory bail applications in PMLA cases. The pre-arrest avenue does not bypass the bail bar. The applicant must still make a prima facie case on both prongs at the pre-arrest stage, often without the benefit of seeing the ECIR. Anticipatory bail in PMLA cases is granted sparingly.

What documents should be attached to a Section 45 PMLA bail application?

Standard annexures include the FIR for the predicate offence, the ECIR or grounds of arrest where available, the arrest memo, all remand orders, the prosecution complaint if filed, co-accused bail orders for parity, medical records if invoking the sick-and-infirm proviso, an affidavit on permanent address, and a list of proposed bail conditions the applicant is willing to undertake. Certified court copies are preferable to family-supplied records.

Which court has jurisdiction over a PMLA bail application: Special Court, Sessions, or High Court?

PMLA bail applications must be filed before the designated Special Court (the Court of Sessions designated under Section 43 PMLA) in the first instance. Direct High Court applications are procedurally improper in most circuits, though some High Courts entertain writ-route bail in exceptional cases (illegal arrest, gross custodial delay). The default workflow is Special Court first, with High Court review on revision or appeal.

How does Section 45 PMLA compare with Section 37 NDPS Act on bail?

Both impose twin conditions: prosecutor opportunity and court satisfaction on reasonable-grounds plus not-likely-to-commit prongs. The textual difference is small. The practical differences are that NDPS bars attract only on commercial-quantity offences (PMLA bars attract on any offence under the Act), and the post-2024 speedy-trial relief is more developed under PMLA than under NDPS where the High Courts remain divided.

How does Section 45 PMLA compare with Section 43D(5) UAPA on bail?

UAPA’s Section 43D(5) applies a single prong test: the court must be of opinion that the accusation is prima facie not true. PMLA’s twin conditions are more demanding because they add the not-likely-to-commit prong. However, UAPA’s standard is harder to meet in practice because there is no proviso architecture, and the speedy-trial relief developed earlier under UAPA (the 2021 K.A. Najeeb ruling) has not yet been fully replicated under PMLA.

What is the difference between bail under PMLA and bail under the BNSS / CrPC general framework?

General BNSS bail under Section 480 of the Bharatiya Nagarik Suraksha Sanhita, 2023 / Section 483 applies the traditional triple test: flight risk, evidence tampering risk, seriousness of the offence. PMLA bail layers the Section 45 twin conditions on top. So a PMLA bail court does two analyses at once: the BNSS triple test and the PMLA twin-conditions test. Both filters must clear before bail can be granted, which is why PMLA bail is materially harder than ordinary non-bailable bail.

Can bail be granted under the PMLA on the ground of delay in trial?

Yes. The post-2024 Supreme Court rulings, particularly Manish Sisodia v. Directorate of Enforcement, 2024 INSC 595, Prem Prakash v. Union of India through Directorate of Enforcement, 2024 INSC 637, and V. Senthil Balaji v. Deputy Director, Directorate of Enforcement, 2024 INSC 739, have read the right to a speedy trial into Section 45. Where the accused has been in custody for an extended period and the trial is not likely to conclude within a reasonable time, constitutional courts have intervened to grant bail. The January 2026 wherewithal ruling sharpened this by adding a prospective prosecution-capacity prong.

What happens if the predicate offence is quashed: does the PMLA case survive?

The PMLA case is fundamentally weakened, though courts have been careful not to say it automatically collapses. The PMLA prosecution requires a scheduled offence at its base. If the predicate is quashed, discharged, or the accused is acquitted of it, the foundational basis for the PMLA charge is shaken. Practitioners increasingly run parallel quashing petitions for the predicate as a strategic route to PMLA bail.

References

Case law

Listed alphabetically by first party name.

• Arvind Dham v. Directorate of Enforcement, 2026 INSC 12. Supreme Court, 6 January 2026. Wherewithal test introduced: prosecution capacity as a bail variable; prolonged pre-trial detention amounts to punishment.
• Directorate of Enforcement v. M. Gopal Reddy, 2023 LiveLaw (SC) 138. Supreme Court, 24 February 2023. Twin conditions of Section 45 apply to anticipatory bail in PMLA cases.
• Kalvakuntla Kavitha v. Directorate of Enforcement, 2024 INSC 632. Supreme Court, 27 August 2024. First proviso to Section 45(1) is not limited to “vulnerable” women; High Courts cannot categorically exclude educated or politically active women from its reach.
• Manish Sisodia v. Directorate of Enforcement, 2024 INSC 595. Supreme Court, 9 August 2024. Right to a speedy trial read into Section 45 PMLA and Section 439 CrPC; bail granted after 17 months in custody.
• Nikesh Tarachand Shah v. Union of India, (2018) 11 SCC 1; AIR 2017 SC 5500. Supreme Court, 23 November 2017. Original Section 45(1) struck down as violative of Articles 14 and 21.
• Pankaj Bansal v. Union of India, 2023 INSC 866. Supreme Court, 3 October 2023. Grounds of arrest under Section 19 PMLA must be furnished in writing.
• Prabir Purkayastha v. State (NCT of Delhi), 2024 INSC 414. Supreme Court, 15 May 2024. Written-grounds-of-arrest requirement extended to UAPA; reaffirms Pankaj Bansal.
• Prem Prakash v. Union of India through Directorate of Enforcement, 2024 INSC 637. Supreme Court, 28 August 2024. “Bail is the rule, jail is the exception” imported into PMLA; prosecution must establish prima facie foundational facts.
• Saumya Chaurasia v. Directorate of Enforcement, 2023 INSC 1073. Supreme Court, 14 December 2023. First proviso to Section 45(1) is discretionary, not mandatory; “may be” interpreted permissively.
• Tarsem Lal v. Directorate of Enforcement Jalandhar Zonal Office, 2024 INSC 434. Supreme Court, 16 May 2024. Twin conditions do not apply where the accused appears on summons after cognisance; ED loses arrest power post-cognisance.
• Union of India v. Kanhaiya Prasad, 2025 INSC 210. Supreme Court, 13 February 2025. Twin conditions are mandatory; bail cannot be granted “casually”; HC bail order set aside.
• V. Senthil Balaji v. Deputy Director, Directorate of Enforcement, 2024 INSC 739. Supreme Court, 26 September 2024. Section 45(1)(ii) does not authorise unreasonably long detention when no possibility of trial concluding within a reasonable time.
• Vijay Madanlal Choudhary v. Union of India, 2022 SCC OnLine SC 929; 2022 LiveLaw (SC) 633. Supreme Court (3-judge bench: Khanwilkar, Maheshwari, Ravikumar, JJ.), 27 July 2022. 2018-amended Section 45 upheld as constitutional; money laundering characterised as a heinous crime.

Statutes

Listed chronologically by year of enactment.

• Unlawful Activities (Prevention) Act, 1967. Section 43D(5) (prima-facie-not-true bail bar for Chapter IV and VI offences).
• Narcotic Drugs and Psychotropic Substances Act, 1985. Section 37 (twin-conditions bail bar for commercial-quantity offences).
• Prevention of Money Laundering Act, 2002. Sections cited: 3, 4, 19, 24, 43, 44, 45 (twin conditions, first proviso, sub-section (1A), sub-section (2)), 50.
• Bharatiya Nagarik Suraksha Sanhita, 2023. Sections cited: 187 (default bail, replacing Section 167 CrPC), 480 (regular bail before Magistrate, replacing Section 437 CrPC), 482 (anticipatory bail, replacing Section 438 CrPC), 483 (bail before Sessions / High Court, replacing Section 439 CrPC).

Further reading

• LiveLaw, reporting and commentary on the August 2024 Supreme Court bail ruling.
• LiveLaw, commentary on the January 2026 wherewithal ruling.
• Bar and Bench, post-2024 quartet analyses.
• SCC OnLine commentary on Section 45 PMLA jurisprudence.

Legal disclaimer

This article is for informational and educational purposes only and does not constitute legal advice. PMLA bail jurisprudence is fact-specific and evolves rapidly; the principles discussed above may not apply identically to any particular case. Readers facing a PMLA proceeding, or advising clients in one, should consult a qualified criminal-litigation advocate with experience in money-laundering matters for case-specific guidance.

Serato DJ Crack 2025Serato DJ PRO Crack