Maintenance Has Three Statutes and No Coherent Doctrine

A wife in India who needs maintenance from a husband has three doors. She can move a Judicial Magistrate of the First Class under section 125 of the Code of Criminal Procedure, now re-enacted as Chapter 9 of the Bharatiya Nagarik Suraksha Sanhita.¹ She can move the same Magistrate under section 20 of the Protection of Women from Domestic Violence Act 2005. If she is Hindu she can move the family court under section 18 of the Hindu Adoptions and Maintenance Act 1956 or, if matrimonial proceedings are pending, sections 24 and 25 of the Hindu Marriage Act 1955.¹² If she is Muslim and divorced she can move the Magistrate under section 3 of the Muslim Women (Protection of Rights on Divorce) Act 1986, the statute that Danial Latifi v Union of India kept alive by reading into it what Parliament had passed it to remove. Three doors that lead to three forums under three statutes with three burdens of proof, three definitions of who counts as a wife, and three theories of what maintenance is for. There is no ordering rule between them. There is no rule even on what happens when she walks through all three.

This is not a system. It is the residue of a series of legislative interventions that responded to particular controversies — Shah Bano, the persistence of bigamous Hindu marriages, the rise of live-in relationships, the inadequacies of section 488 of the 1898 Code — each by adding another track, never by integrating the existing ones. The Supreme Court tried, in 2020, to make the residue work. In Rajnesh v Neha the Court issued binding directions: a uniform Affidavit of Disclosure of Assets and Liabilities for every maintenance application in India, criteria for fixing quantum, mandatory payment from the date of application, and a disclosure-and-set-off rule under which an applicant pursuing maintenance under one statute must reveal what she has already obtained or is pursuing under any other.¹⁰ Rajnesh is a careful attempt at procedural coordination. It is also a confession. The Court is doing through directions what the legislature has refused to do through code — coordinate the substantive law. And procedural coordination cannot, in the end, do what only substantive law can: decide what maintenance is for, who is entitled to it, and on what theory.

Three statutes, three theories

The three statutes do not just have different procedural settings. They proceed from different theories of what maintenance is. The Supreme Court has long treated Chapter IX of the CrPC as 'a self-contained code' for the criminal-procedure remedy.² The civil-law remedies under the Hindu statutes have their own self-contained character. The DV Act creates its own. None of the three was written with the others in view, and the doctrinal consequences flow from that fact.

Section 125 CrPC, in its origin and in its judicial reading, is a vagrancy-prevention statute. Bhagwan v Kamla (1974) and the long line that follows treat it as a 'summary remedy to save dependants from destitution and vagrancy', operating 'apart from and independent of' personal law.³ The Magistrate's jurisdiction is 'preventive, rather than remedial or punitive'. The applicant must be 'unable to maintain herself' — a condition expressly inserted into the 1973 Code to fix what was thought to be an over-reach of the older section 488. The right is statutory; it survives even where personal law would deny it; but the quantum is calibrated to the bare minimum a person needs to avoid destitution. The provision is, in the texture of its reasoning, a public-order statute that happens to operate inside family disputes.

The Hindu Marriage Act and the Hindu Adoptions and Maintenance Act proceed from an entirely different premise. Sections 24 and 25 of the HMA, and sections 18 and 20 of the HAMA, are status statutes. They flow from the existence of the marital relationship. They do not require the applicant to plead destitution. Section 18 of the HAMA entitles a Hindu wife to maintenance 'during her lifetime', calibrated not to a survival minimum but to her standard of living in the matrimonial home. The remedy is in the civil or family court, not summary; the determination is final, not preventive. The theory is one of marital obligation, not destitution-prevention.

The Domestic Violence Act 2005 sits on a third theory altogether. It does not depend on the existence of a marriage at all. Section 2(f) defines a 'domestic relationship' to include relationships 'in the nature of marriage'; section 2(s) defines a shared household; section 3 defines domestic violence to include 'economic abuse'. The Act's maintenance order under section 20 is not framed as relief from destitution and not framed as marital obligation but as a remedy for a wrong — economic abuse — that exists whether the relationship is marital or not. D Velusamy v D Patchaiammal (2010) made the doctrinal architecture of this third track explicit. The Court held that a 'relationship in the nature of marriage' under the 2005 Act is 'akin to a common law marriage' and requires that the couple hold themselves out as spouses, be of legal age, be otherwise qualified to marry, and have voluntarily cohabited in a shared household for a significant period.⁵ The Court was careful to add that 'not all live-in relationships will amount to a relationship in the nature of marriage' — a 'keep' maintained financially and used for sexual purposes would not count.

Three theories. Vagrancy-prevention, marital obligation, remedy for economic abuse. They are not subsets of each other. They do not even share a common definition of who the relevant relationship is between.

Who counts as a wife

Nothing illustrates the fragmentation better than the law on who is entitled to invoke each statute. The threshold question — is the applicant a 'wife' — gets a different answer in each forum, and within section 125 CrPC it gets several different answers depending on which line of authority the Magistrate follows.

The textual rule in section 125 is that 'wife' includes a woman divorced or who has obtained a divorce and has not remarried (Explanation (b) to sub-section (1)). The Yamunabai-Savitaben line, decided by the Supreme Court itself, reads this strictly: a second wife whose marriage is void because the first marriage subsisted is not a legally wedded wife and cannot claim under section 125.⁷ The Badshah line, also of the Supreme Court, reads it purposively: where the first marriage was concealed from the second wife, she will be 'treated as the wife' for section 125 purposes. Badshah v Sou Urmila Badshah Godse is candid about what it is doing — Sikri J writes that 'while interpreting a statute the court may not only take into consideration the purpose for which the statute was enacted, but also the mischief it seeks to suppress'.⁶ The 'mischief' here is that the husband would otherwise be 'defrauding the wife'. The reasoning is moral and the reasoning is correct. It is also a substitution of judicial purpose for legislative text.

Both lines are good Supreme Court law. The Court in Chanmuniya v Virendra Kumar Singh Kushwaha (2011) noticed the contradiction and referred to a larger Bench three questions, including whether a long-term cohabitation should be presumed to be marriage for section 125 purposes.¹¹ The reference is, fifteen years on, still pending. In the meantime, Magistrates apply Yamunabai or Badshah depending on the facts and their reading of what the spirit of section 125 requires. This is not jurisprudence in disagreement with itself. It is a statute that has had three different things asked of it and has been read three different ways.

The DV Act compounds the picture. Velusamy draws a line between a 'relationship in the nature of marriage' and a 'live-in relationship' that does not qualify. The line is administrable on paper and impossible to administer in practice. The Court asks whether the parties held themselves out as spouses, were of legal age and otherwise qualified to marry, voluntarily cohabited, and lived in a shared household for a 'significant period'. None of these criteria is sharper than 'significant'. The Magistrate has to make findings of social fact about how a couple presented themselves to the world. The findings will turn on local mores, on the kind of evidence parties choose to lead, and ultimately on the Magistrate's intuition about whether this is the kind of relationship the statute meant to protect. In a court system that processes hundreds of maintenance applications per Magistrate per year, this is not a workable threshold rule. It is a remand-prone one.

A divorced Muslim woman is in yet another position. Shah Bano brought her within section 125 CrPC.⁸ The Muslim Women (Protection of Rights on Divorce) Act 1986 was passed to put her outside it. Danial Latifi v Union of India read the 1986 Act so that, by the time the Constitution Bench was done with it, the Act 'actually and in reality codifies what was stated in Shah Bano case'.⁹ The husband's obligation under section 3(1)(a) of the 1986 Act is to make a 'reasonable and fair provision and maintenance' that extends, on the Court's reading, to the whole of the divorced wife's life unless she remarries. The reasoning of the judgment is, by its own admission, a constitutional-avoidance manoeuvre: a narrower reading would render the Act discriminatory under Articles 14, 15 and 21 and the Court is unwilling to so hold. The result is a special statutory route for divorced Muslim women that, after sustained judicial work, produces an outcome substantively similar to section 125 CrPC — but procedurally distinct, with the option under section 5 of the 1986 Act for the parties jointly to elect into Chapter IX of the Code instead. The procedure-substance gap is enormous.

The unspoken coordination rule: res judicata, set-off, and judicial discretion

The three tracks coexist, but they do not coordinate. The closest thing to a coordination rule is the cluster of judge-made principles that have grown around the avoidance of double recovery.

The Supreme Court in Nanak Chand v Chandra Kishore Aggarwal (1970) held that section 125 CrPC and section 24 of the HMA are 'separate and independent' and 'the scope is different' — relief under one does not bar the other.⁴ The High Courts have applied this to multiple combinations: section 125 alongside section 18 HAMA, section 125 alongside section 20 of the DV Act, section 24 HMA alongside section 20 DV Act. The settled rule is that all are independent and may run in parallel; the unsettled rule is what to do when they do.

What happens is a patchwork. Some courts kept the section 125 order 'in abeyance' while pendente lite under section 24 HMA operates (see Kamlesh Kumari v Aman Kishore). Others awarded under both and adjusted quantum to avoid 'double recovery'. The principle that emerges is: pay each independently determined sum, but set off any amount actually received under one against the obligation under another. Rajnesh v Neha formalised this into a direction. Every applicant must disclose existing or pending maintenance orders. Every court must adjust quantum to avoid double recovery. The result is logical from an accounting perspective and unprincipled from a doctrinal one. We have, in effect, said: the substantive entitlements differ, the burdens differ, the theories differ — but at the end of the day the money is just money, so we will tot it up and pretend the difference is administrative.

The disclosure-and-set-off rule treats the three statutes as parallel funding lines for a single underlying claim. They are not. The claim under section 125 is a destitution-prevention claim. The claim under section 18 HAMA is a status-based marital-obligation claim. The claim under section 20 of the DV Act is a remedy-for-wrong claim. If you allow set-off across them, you have collapsed three substantive theories into one accounting fund. That is not what the legislature did. That is what the Court has had to do because the legislature did not coordinate the statutes when it passed them.

What Rajnesh actually does — and what it cannot do

The directions in Rajnesh v Neha are useful. The Affidavit of Disclosure of Assets and Liabilities is a real procedural advance. The requirement that maintenance be paid from the date of application, not the date of order, fixes a long-standing injustice. The disclosure obligation across statutes prevents the most egregious double-recovery cases.

But what Rajnesh cannot do is the substantive work. It cannot decide what maintenance is for. It cannot decide whether a divorced second wife, married in ignorance of a first marriage, is a 'wife' for section 125 purposes. It cannot decide whether a long-term cohabitant qualifies under section 2(f) of the DV Act. It cannot decide whether the section 3(1)(a) provision for a divorced Muslim woman is co-extensive with section 125 CrPC or operates on a different theory. These are substantive questions and they require substantive answers, not affidavits.

What Rajnesh can do, and does, is impose uniformity on the second-order question: given that an applicant has been found entitled under one or more of the three statutes, what is the quantum and how is it coordinated? On that question its directions are good. But the first-order question — is she entitled — is asked three different ways in three different statutes and is answered differently by different lines of Supreme Court authority. The procedural uniformity of Rajnesh sits on top of substantive disorder.

The counter-argument, addressed honestly

The honest version of the case for the present system runs like this. Indian family law is, by constitutional choice, plural. Article 44 of the Constitution promises a Uniform Civil Code, but no government in seventy-five years has tried to legislate one for marriage and divorce. The personal laws of Hindus, Muslims, Christians and Parsis remain on the books. Section 125 CrPC was designed precisely to operate above this plurality — to give a secular, summary, destitution-prevention remedy regardless of which personal-law regime the parties belong to. The DV Act extended the principle to a wider class of women, including those whom personal law would not recognise as wives at all. Each track exists because each does something the others do not. To insist on a single coordinating doctrine is to require, in effect, what India has refused to legislate.

This argument has real weight. The fragmentation is not pure legislative carelessness. It is the price of keeping a secular criminal-procedure remedy alongside three sets of personal-law civil remedies. A divorced Muslim woman who could not, in 1986, get past section 127(3)(b) on the Shah Bano facts would have been left without a remedy if section 125 were unavailable and the Muslim personal-law iddat rule applied without qualification. The 1986 Act was a political compromise; the Constitution Bench in Danial Latifi read it in a way that preserved the substantive protection that Shah Bano had recognised. Without parallel tracks, that route would not have been available. The same is true of the DV Act: a cohabiting woman whose relationship the personal law does not recognise needs a track that does not depend on personal-law marital status. Section 125 alone might have done it, with sufficient judicial gymnastics on the meaning of 'wife', but the DV Act gave it cleaner doctrinal architecture. The tracks exist because the underlying entitlements they protect are not identical, and forcing them into a single doctrine would have meant choosing one of the three theories and losing the protections that the other two provide.

I accept this. The argument explains why we have three statutes. It does not explain why we have no ordering rule between them. Plurality of substantive law does not require absence of coordination. The Civil Procedure Code coordinates with the Family Courts Act; the Limitation Act coordinates with both; the Indian Evidence Act applies across all civil proceedings. The example of section 5 of the 1986 Muslim Women Act — letting the parties jointly elect into Chapter IX of the CrPC — shows that statutory coordination is possible. The legislature could have written a coordination chapter into the DV Act when it passed it in 2005, specifying when section 20 DV Act ousts section 125 CrPC, when it runs in addition, and when the personal-law civil remedies must be elected over the DV Act remedy. It did not. The choice was to leave coordination to the courts, and the courts have had to coordinate through accounting and discretion. The alternative is not uniform civil code. It is statutory coordination of multiple regimes — a much more modest legislative task that has not been attempted.

What the criminal-procedure track was designed to do, and what it now does

It is worth pausing on what section 125 was for, because the statute has drifted a long way from its origins and the drift is part of what makes coordination so hard. Section 488 of the 1898 Code was a preventive measure aimed at vagrancy. The Magistrate could pass a summary order against a man with sufficient means who refused or neglected to maintain his wife, child or parent. The amount was capped at five hundred rupees a month. The proceeding was preventive, not punitive, and the Magistrate had no power to determine the substantive matrimonial rights of the parties. Any order he passed could be re-litigated in a properly framed civil suit. That was the system: a fast, capped, preventive order in the criminal court; the substantive determination, if anyone wanted it, in the civil court.

The 1973 Code preserved this design but inserted the 'unable to maintain herself' threshold and the requirement that a married daughter who had attained majority be excluded. The Statement of Objects and Reasons of the 2001 Amendment Act records the rationale for the changes Parliament made next: the cap had been set in 1958 and had become absurd, and the proceedings were taking years to dispose of so an interim-maintenance power was inserted. Parliament deleted the cap. The Supreme Court in Manoj Yadav v Pushpa held that the state amendments to section 125 fixing rupee ceilings became invalid because they were inconsistent with the parliamentary deletion of the central ceiling. The provision, in 2001, ceased to be a capped preventive remedy and became an uncapped support order with judicial discretion as to quantum.

What the uncapped section 125 became, in practice, is a substantive maintenance regime in everything but name. Chaturbhuj v Sita Bai made the test explicit: the quantum must let the wife 'maintain a standard of living, neither luxurious nor penurious but what is consistent with the status of the family'. That is not a vagrancy-prevention test. That is a marital-obligation test. The provision continues to describe itself as preventive and to require the destitution threshold at the entry gate, but the quantum at the exit gate is calibrated to marital standard of living. The doctrinal apparatus of Bhagwan v Kamla sits uneasily on top of the post-2001 quantum jurisprudence.

This drift matters for the coordination question because the doctrinal distinction between section 125 and section 18 HAMA has collapsed at the quantum end while remaining intact at the entry-threshold end. The Magistrate hearing a section 125 case asks one set of questions to decide whether the wife is entitled — 'unable to maintain herself' — and a different set to decide how much she gets — 'standard of family living'. The first set is preventive, the second is substantive. The civil-law tracks ask the substantive question at both ends. The DV Act asks neither. The result is that an entitled applicant gets, under section 125, an amount that is broadly the same as what she would have got under the civil-law statutes — but only after clearing a destitution-threshold gate that the civil-law statutes do not impose. The gate is doing different work than the quantum, and the statute does not say which is governing.

What an honest fix would look like

The substantive law of maintenance does not need a single theory; it needs three clear ones and an ordering rule between them. The legislature could, without disturbing the personal-law systems, do four things.

First, decide whether section 125 CrPC is a destitution-prevention floor or a substantive entitlement. The 1973 'unable to maintain herself' threshold, read with the 2001 abolition of the Rs 500 ceiling, has produced a provision that operates as both — a floor in some cases and a substantive maintenance award in others. The Yamunabai-Badshah split tracks this confusion. If section 125 is a floor, the Yamunabai line is correct: only legally wedded wives qualify, the provision protects against destitution, and an aggrieved second wife must seek her remedy elsewhere (civil suit, DV Act, criminal complaint for fraudulent marriage). If section 125 is a substantive entitlement, Badshah is correct, but then the legislature should say so by amending Explanation (b) to include a woman whose putative marriage was procured by concealment of a subsisting marriage. Leaving the question to a Bench-by-Bench reading of 'mischief' is unworkable.

Second, define 'relationship in the nature of marriage' under section 2(f) of the DV Act. Velusamy's common-law-marriage analogy is the best the Court can do with the existing text. The text is the problem. Parliament should specify the durational and conduct-based criteria. It need not extend the definition to every cohabitation, and the policy choice on where to draw the line is a legitimate legislative one — but the line should be drawn legislatively, not litigated repeatedly.

Third, write a coordination chapter. Specify whether a section 20 DV Act order is in addition to, in lieu of, or set-off against a section 125 CrPC order or a section 18 HAMA decree. Specify whether a Muslim woman entitled under section 3(1)(a) of the 1986 Act can simultaneously claim under section 125. Specify whether a Hindu wife who has obtained pendente lite under section 24 HMA can take a section 125 order in parallel, and what becomes of either when permanent alimony is decreed under section 25. The Rajnesh set-off principle is good practice and bad doctrine. It should be statutory, with the substantive question of which entitlement is primary expressly decided.

Fourth, fix the Chanmuniya reference. Either the Supreme Court decides the question pending before the larger Bench, or the legislature pre-empts it. Fifteen years of pending reference on whether long-term cohabitation entitles a woman to section 125 maintenance is not delay — it is institutional refusal to answer a question that the legal profession knows the answer to in some courts and not others.

Why the system stays as it is

None of this will happen because each of the three tracks has a constituency that benefits from its survival on the current terms. The personal-law civil bar likes section 18 HAMA, sections 24 and 25 HMA, and the 1986 Act because they keep family-law work in the family court and produce final, appealable orders. The criminal-defence bar likes section 125 because its summary character and the lower burden of proof on the wife produces fast pendente-lite money that the husband can be made to pay even before the personal-law civil case has been pleaded. The women's-rights bar likes the DV Act because section 20 sits inside a wider relief structure (protection orders, residence orders, custody orders) and does not require the applicant to call herself a wife at all. Each constituency has structured its practice around its preferred track. The Magistrate at the bottom of all three has structured her docket around running them in parallel. Rajnesh made that running-in-parallel a little more orderly. It did not make it any more coherent.

So the system stays. A destitute wife in a small town finds the lawyer she can afford, the lawyer files in the forum he knows best, and the substantive theory of her claim is whichever theory that forum applies. If the case settles, the set-off rule from Rajnesh kicks in. If it does not, she gets a different quantum, on a different timeline, with a different finality, depending on which door she walked through.

The doctrinal point is that this is not maintenance law. It is three different things that go by the same name, with an accounting rule sitting on top. The honest reading of Rajnesh v Neha is that the Supreme Court has done the only thing left to do — manage the residue procedurally — because the legislative work of coordinating the substantive law has not been done. The Court can manage indefinitely. It cannot replace the legislature in deciding what maintenance in India is for, and that is the question the system as a whole has refused to answer.

Get the week's biggest decisions, distilled.

A weekly digest for lawyers, students and the legally curious. Every Friday morning. Free.