The ordinance is the point at which the Indian Constitution lets the executive make law. Article 123 empowers the President, when Parliament is not in session and circumstances require immediate action, to promulgate an instrument that carries the same force and effect as an Act of Parliament. Article 213 confers the corresponding power on a Governor in relation to a State legislature, on narrower terms. Almost everything written about the subject in public discussion captures a part of this arrangement and misses the discipline of the whole. The six weeks that dominate the commentary are counted from the wrong event. An emergency is treated as the condition when the text imposes none. The three restrictions that separate the Governor’s power from the President’s are listed rarely, though they are among the most examinable propositions in Indian constitutional law. This article states the power exactly: its two textual sources, the condition on which it may be exercised, the arithmetic that bounds its life, the limits on its content, the identity of the authority that in substance wields it, and the line of decisions through which the Supreme Court has fixed what the power is and what it is not.

The One Test for this pillar is easily stated. A reader who has finished should be able to say that an ordinance is legislation made by the executive when the legislature is not sitting; to state the conditions on which it may be made and the time limits that govern its life; and to name the three respects in which the Governor’s power is narrower than the President’s. Everything in the pages that follow serves that test. The companion pillar in this series takes the leading judgment on re-promulgation and on the fate of acts done under an instrument that has ceased to operate, and develops it in full. This article’s territory is the power itself: where it comes from, how it is triggered, how long its product endures, what its product may contain, and how the courts have characterised it across five decades. Where the two territories meet, this article states the holding at headline level and leaves the development to its owner, so that the pair cannot be made to conflict.

The subject repays precision because the power is used constantly. Ordinances are not a curiosity of the early Republic that faded as parliamentary practice matured. They have been promulgated in substantial numbers in every decade since the Constitution came into force, and the count kept by PRS Legislative Research records 637 instruments at the Centre between the first Lok Sabha in 1952 and the year 2014, an average of more than ten a year. A device used ten times a year by governments of every complexion is a standing feature of the legislative system, and it deserves to be described with the same care that the series gives to an Act. It also deserves to be described without the two polemical shortcuts that dominate discussion of it. The first shortcut treats every ordinance as an executive usurpation, as though the Constitution had not itself created the power and hedged it with conditions. The second treats every criticism as obstruction, as though the conditions were decorative. The text supports neither shortcut, and the case law supports neither. What the text and the cases together describe is a real legislative power, deliberately conferred, bounded by a condition the executive itself controls, by a clock the legislature can stop early, and by a content limit that ties the instrument to the competence of the legislature whose place it temporarily takes.

Where the power sits in the Constitution

Article 123 stands in Chapter III of Part V of the Constitution, a chapter of three articles under the heading Legislative Powers of the President. Article 213 stands in Chapter IV of Part VI, headed Legislative Power of the Governor. The placement is not ornamental. When the Supreme Court came to characterise the power in A. K. Roy v Union of India in 1982, the chapter headings formed part of the reasoning by which the Court concluded that the President, in promulgating an ordinance, exercises legislative and not executive power. The heading announces the conclusion that the case law then worked out: the instrument is law, made by an authority other than the legislature, under a power the Constitution itself classifies as legislative.

The lineage of the device runs behind the Constitution. The Indian Councils Act of 1861 armed the Governor-General with a power to make ordinances in cases of emergency, with a life of six months. The Government of India Act of 1935 placed the power in section 42 for the Governor-General and in section 88 for a provincial Governor, and the six weeks from reassembly that the Constitution prescribes is a descendant of the arrangement the 1935 Act worked with. The Constituent Assembly debated the inheritance and kept it, over objections that are still quoted. Members warned that a power to legislate without the legislature would be used for convenience rather than necessity. H. V. Kamath was among those who pressed the Assembly on the risk. Dr B. R. Ambedkar, moving the draft provisions, defended the power on the ground that a legislature cannot sit continuously and that circumstances may arise which cannot wait for the next session: the mechanism was, in his presentation, an exceptional recourse for a gap in the legislative calendar, not a parallel legislature. The Assembly adopted the power with the safeguards that the text carries, and the tension between the defence and the warning has structured every later debate, judicial and political, on the subject.

Why does the Constitution place the ordinance power among the legislative powers?

The chapter headings are the first reason. Article 123 is placed under the heading Legislative Powers of the President, and Article 213 under Legislative Power of the Governor, and the Supreme Court in A. K. Roy treated that placement as a pointer to the character of the power.

The President and the Governor, when they promulgate, are exercising a legislative function conferred by the Constitution, and the product of the exercise is law in the full constitutional sense, not an administrative direction dressed in statutory form.

The characterisation does real work. Because the power is legislative, its product is law for the purposes of Article 13, which forbids the State from making law that abridges fundamental rights, and an ordinance may therefore be struck down for violating Part III exactly as an Act may be. Because the power is legislative, its product is law for the purposes of Article 21, as A. K. Roy held, so that a deprivation of life or personal liberty under an ordinance must answer the same constitutional tests as a deprivation under an Act. Because the power is legislative, the motives and the quality of the deliberation behind it are not examinable in the way the motives behind an executive order are examinable, a proposition the Court worked out in T. Venkata Reddy v State of Andhra Pradesh in 1985. And because the power is legislative, its limits are the limits of legislation: the distribution of legislative competence between the Union and the States, the fundamental rights, and every other constitutional restriction on law-making. The one great qualification the characterisation carries is the satisfaction on which the power is conditioned, and the extent to which a court may look behind that satisfaction. That question runs through the case law from 1970 onwards and receives its own treatment below.

The constitutional neighbourhood of the power also includes the provisions that govern sessions. Article 85 empowers the President to summon each House of Parliament, to prorogue the Houses or either House, and to dissolve the House of the People; its proviso states that six months shall not intervene between the last sitting of a House in one session and the date appointed for its first sitting in the next session. Article 174 makes the corresponding provision for a State legislature, with the same six-month outer limit between sessions. These provisions matter to the ordinance because they fix the longest gap the calendar can contain, and therefore the longest period for which the condition of the power, that the legislature is not in session, can continuously hold. They also matter because the authority that advises on summoning and prorogation is the same Council of Ministers on whose advice an ordinance is promulgated, a structural feature of the arrangement to which this article returns, since it is the feature on which the deepest criticism of the power rests.

The text of Article 123, clause by clause

Article 123 has three clauses, and each carries a distinct part of the discipline. The first confers the power and states its two conditions. The second gives the product its force and fixes its fate. The third states the limit on its content. Public discussion habitually compresses the three into a single impression, and the compression is where the errors enter.

The first clause provides that if at any time, except when both Houses of Parliament are in session, the President is satisfied that circumstances exist which render it necessary for him to take immediate action, he may promulgate such ordinances as the circumstances appear to him to require. Three features of the sentence deserve emphasis. The first is the shape of the recess condition. The words are not that Parliament must be wholly out of session. The bar operates only when both Houses are in session, so that the power is available whenever at least one House is not sitting. A session of the Rajya Sabha alone does not close the power, and a session of the Lok Sabha alone does not close it. Textbooks that state the condition as Parliament not being in session are giving the practical norm, since the Houses ordinarily sit together, but the text is wider than the norm, and the width matters when the Houses are summoned on different dates. The second feature is the satisfaction. The President must be satisfied that circumstances exist which render immediate action necessary. The text requires circumstances and necessity; it does not require an emergency, a war, a breakdown, or any named category of crisis, and the absence of the word emergency from the clause is the answer to the commonest misdescription of the power, which treats it as an emergency provision. It stands in Part V, among the ordinary legislative provisions, not in Part XVIII with the emergency provisions, and the satisfaction it demands is a satisfaction about the need for immediate legislative action, not about the existence of a proclaimed crisis. The third feature is the tailoring of the product to the occasion: the President may promulgate such ordinances as the circumstances appear to him to require. The instrument is in form responsive to the circumstances that justify it, though the courts have not converted that responsiveness into a test of proportionality, and the width of the satisfaction is examined in the case law treatment below.

The second clause provides that an ordinance promulgated under the article shall have the same force and effect as an Act of Parliament, but that every such ordinance shall be laid before both Houses of Parliament and shall cease to operate at the expiration of six weeks from the reassembly of Parliament, or, if before the expiration of that period resolutions disapproving it are passed by both Houses, upon the passing of the second of those resolutions; and that it may be withdrawn at any time by the President. An Explanation adds that where the Houses of Parliament are summoned to reassemble on different dates, the period of six weeks shall be reckoned from the later of those dates. Four legal consequences are packed into the clause. The force and effect are plenary from the moment of promulgation: the instrument is not a draft, a proposal, or a provisional direction awaiting ratification, but operative law that courts enforce, administrations act under, and citizens must obey, from the day it is promulgated. The laying requirement subjects the instrument to the legislature’s attention as a constitutional obligation, not a courtesy. The cessation rule fixes the outer term of the instrument’s life by reference to reassembly, and the disapproval rule gives the legislature a way to end the life earlier by resolution, without the need to pass a repealing Act. The withdrawal power lets the promulgating authority end the instrument at any time, including before Parliament reassembles, which is the route taken when a government decides that a measure should not proceed or should proceed in altered form through a Bill.

The third clause provides that if and so far as an ordinance makes any provision which Parliament would not under the Constitution be competent to enact, it shall be void. The limit is stated as a competence rule and operates provision by provision: the offending provision is void to the extent of the incompetence. Its content is worked out in the treatment of the subject’s limits below; its importance in the architecture is that it ties the executive’s legislative product to the same distribution of powers that binds the legislature, so that the ordinance route cannot enlarge the Union’s legislative field by one entry.

The ordinance clock: six months, six weeks, and the arithmetic of a maximum life

The life of an ordinance is governed by two periods drawn from two different sources, and the failure to separate them is the source of the commonest error in the literature. The first period is the maximum gap between sessions, fixed by the proviso to Article 85 at six months: six months shall not intervene between the last sitting of a House in one session and the first sitting in the next. The second period is the six weeks of Article 123(2), which runs not from promulgation but from the reassembly of Parliament. An ordinance promulgated on the first day on which the power becomes available, in a case where the government then waits the longest lawful interval before summoning Parliament, can therefore live for the six months of the inter-session gap and then for the six weeks that follow reassembly: seven and a half months at the outside. The Supreme Court in A. K. Roy stated the maximum in exactly those terms, describing the difference between an Act and an ordinance as a difference of duration, the ordinance ceasing to exist after seven and a half months of its promulgation at the very most unless disapproved earlier.

The six weeks deserve a closer statement, because three sub-rules hide inside them. The first is that the period runs from reassembly, and where the two Houses reassemble on different dates the Explanation moves the start to the later date, so the instrument cannot be shortened by a staggered recall. The second is that the period is a period of the reassembly session’s calendar: the ordinance must be laid before both Houses, and the six weeks are the window within which the legislature may disapprove, within which the government ordinarily introduces and carries the replacing Bill, and at whose end the instrument ceases to operate of its own force if nothing has been done. The third is that cessation at the end of the six weeks is automatic and requires no act of any authority: the instrument does not lapse because someone declares it lapsed, and it does not survive because no one remembered it. It ceases to operate by the force of the constitutional text.

Disapproval is the legislature’s early stop on the clock. A resolution disapproving the ordinance must be passed by both Houses, and the instrument ceases to operate upon the passing of the second of the resolutions. The mechanism is deliberately lighter than repeal: no Bill, no assent, no third reading choreography, only a resolution of each House within the six weeks. Its practical significance is that a government commanding a majority in the Lok Sabha but not in the Rajya Sabha cannot treat survival as assured, since the second resolution may come from a House it does not control. Withdrawal is the executive’s own early stop, available at any time, and exercised where the government prefers to drop the measure, to replace it with a different measure, or to concede the political argument and take the subject back to the drafting table. The fourth outcome, replacement by an Act, is the ordinary intended path: a Bill in the terms of the ordinance, or in terms modified by debate, is passed by both Houses and assented to, and the law continues without a break, the ordinance having served as a bridge across the recess.

The comparison that examiners and commentators reach for is with the Governor-General’s ordinance under the Government of India Act of 1935, which ran for six months from promulgation unless earlier disapproved or withdrawn, a single fixed term rather than the Constitution’s two-stage clock. The constitutional design is more favourable to legislative control in one respect and less in another. It is more favourable because the operative period after the legislature returns is short and tied to the session, so the legislature confronts the instrument promptly. It is less favourable because the inter-session gap, which can stretch the total life to seven and a half months, is bounded by a six-month rule that the government of the day can use to the full, since the timing of summoning within that bound is a matter of executive advice.

The findable artifact: the ordinance clock as a step table

The sequence is set out below as a single table, because the discipline of the power is a discipline of steps and dates, and a reader who can follow the table can test any account of any particular ordinance against it. The final column records the position of acts done under the instrument at each outcome, at the level of principle this pillar states; the developed doctrine on that question belongs to the companion article on the re-promulgation judgment and is flagged in the table rather than resolved in it.

Step	Event	Period or trigger	Consequence for the ordinance	Position of acts done under it
1	Decision to legislate during a recess	The Council of Ministers advises promulgation	No legal effect yet; the satisfaction is formed	No acts arise
2	Condition verified	At least one House of Parliament is not in session	The power under Article 123 becomes available	No acts arise
3	Promulgation	Publication of the instrument; takes effect at once unless a later commencement is stated	Operative law with the force of an Act of Parliament	Acts done from this point are done under valid law
4	Effect on existing law	The instrument amends, repeals, or supplements the law in its field	The amended position governs while the instrument operates	Rights and liabilities accrue under the amended position
5	Reassembly of Parliament	Summoned within six months of the last sitting of the previous session	The six-week period of Article 123(2) begins, from the later date if the Houses return on different dates	The accrual of rights and liabilities continues
6	Laying before both Houses	On reassembly	The legislature is seized of the instrument as a constitutional obligation	No change
7	Outcome one: replacement	A Bill in the terms of the instrument is passed and assented to within the six weeks	The Act continues the law without a break	Acts done are carried into the Act’s regime
8	Outcome two: expiry	The six weeks run out with no replacing Act	The instrument ceases to operate of its own force	Whether completed acts endure is the question the companion article answers
9	Outcome three: disapproval	Resolutions of disapproval passed by both Houses, the second passage ending the life	The instrument ceases to operate at once	The same question arises, sharpened by the legislature’s rejection
10	Outcome four: withdrawal	The President withdraws the instrument at any time	The instrument ends from withdrawal	Acts done before withdrawal stand on the law as it operated while in force

The table exposes the feature that gives the subject its difficulty. Every outcome except replacement leaves a residue: things were done, money was collected or spent, appointments were made, prosecutions were begun, posts were abolished, under an instrument that has ended. The text of Article 123 says the instrument ceases to operate; it does not say that the past is rewritten. The distance between those two propositions is the distance the case law on effects has had to travel, and the companion article on the leading judgment travels it.

Article 213 and the Governor’s narrower power

The State provision follows the Union provision closely and departs from it at points that are deliberate. Article 213(1) empowers the Governor, if at any time, except when the Legislative Assembly of the State is in session, or where there is a Legislative Council in the State, except when both Houses of the Legislature are in session, he is satisfied that circumstances exist which render it necessary for him to take immediate action, to promulgate such ordinances as the circumstances appear to him to require. Article 213(2) gives the product the same force and effect as an Act of the Legislature of the State assented to by the Governor, requires it to be laid before the Legislative Assembly or, where there is a Council, before both Houses, and makes it cease to operate at the expiration of six weeks from the reassembly of the Legislature, or earlier if a resolution disapproving it is passed by the Legislative Assembly and agreed to by the Legislative Council, if any; it may likewise be withdrawn by the Governor at any time. Article 213(3) applies the competence limit in State terms: a provision that the State legislature would not be competent to enact is void. The differences from Article 123 in the machinery are two, and both are minor. In a unicameral State the disapproval of the single House suffices, since there is no Council to agree. In a bicameral State the disapproving resolution must originate in the Assembly and be agreed to by the Council, an asymmetry that reflects the Assembly’s primacy in the State scheme. The great differences are not in the machinery. They are the three restrictions in the proviso to Article 213(1), which have no counterpart at the Union level and which every account of the subject is required to state exactly.

The proviso states that the Governor shall not, without instructions from the President, promulgate any such ordinance if, first, a Bill containing the same provisions would under the Constitution have required the previous sanction of the President for the introduction thereof into the Legislature; or, second, he would have deemed it necessary to reserve a Bill containing the same provisions for the consideration of the President; or, third, an Act of the Legislature of the State containing the same provisions would under the Constitution have been invalid unless, having been reserved for the consideration of the President, it had received the assent of the President. The three cases are best understood as three doors to the same room, which is the room of Union supervision over State legislation in the sensitive fields. The first door is prior sanction: some Bills, of which the standard example is a Bill that would impose restrictions on trade and commerce of the kind Article 304 contemplates, cannot even be introduced into a State legislature without the President’s previous sanction, and an ordinance carrying the same provisions cannot be used to slip past the introduction gate. The second door is reservation in the Governor’s own judgment: if a Bill in those terms were before him for assent and he would have deemed it necessary to reserve it for the President, he cannot take the same content by the ordinance route without instructions, since that would let the reservation judgment be defeated by a change of instrument. The third door is invalidity curable by reservation and assent: some State enactments, paradigmatically a law repugnant to a Union law on a Concurrent List subject, are invalid unless reserved and assented to, which is the arrangement Article 254’s treatment of repugnancy works out in detail; an ordinance in those terms, promulgated without instructions, would manufacture validity the legislative route could not have produced on its own.

The reason for the asymmetry between the two articles is structural. The President is the head of the Union executive, and there is no higher executive authority within the constitutional scheme to instruct him; a restriction of the State kind would have no one to operate it. The Governor, by contrast, sits inside a federal scheme in which the Union has defined interests in defined fields of State legislation, interests the Constitution protects through the sanction, reservation, and assent machinery of the ordinary legislative process. If the ordinance route bypassed that machinery, the protections would exist only for governments polite enough to proceed by Bill. The proviso therefore reads the ordinance power down to the level of the legislative power as the federal scheme conditions it: a Governor can do by ordinance, without instructions, only what the State legislature could have done by Act without engaging the President’s role. Where the President’s role would have been engaged, it is engaged still, and the instructions are its form.

The instructions device also changes the politics of the State power in a way worth stating without embellishment. A Governor who needs instructions must disclose the proposed content to the Union government and await its decision, so the Union acquires a preview and a veto over a defined class of State ordinances before they exist. Critics of the arrangement, among them the Sarkaria Commission in its examination of Centre-State relations, treated the Governor’s ordinance power as one to be used sparingly and in exceptional circumstances, and the Commission’s caution reflects the double character of the State instrument: it is both an exceptional legislative device and a point at which Union supervision touches State law-making before the fact. Defenders of the proviso answer that the three cases are narrow and defined, that they track fields where the Constitution itself has already qualified State autonomy, and that a State government with an ordinary subject and an ordinary urgency faces none of them. Both statements are consistent with the text; the text itself confines the proviso to the three cases, and neither statement licenses a Governor to seek instructions, or the Union to demand them, outside those cases.

One further difference between the articles is a matter of the satisfaction’s setting rather than of text. The Governor acts on the aid and advice of the State Council of Ministers in promulgating an ordinance, as the President acts on the aid and advice of the Union Council of Ministers, and the instrument is in substance the State government’s legislation. But the Governor’s position carries a complication the President’s does not: the same Governor who promulgates on ministerial advice in this field is the authority who, in the reserved-Bill cases, exercises a judgment the Constitution frames as his own, whether he would have deemed it necessary to reserve a Bill. The proviso therefore forces the Governor’s office to hold two postures at once, the advised posture of the promulgating authority and the judgment posture of the reserving authority, and the friction between them has made Article 213 the more litigated and the more politically charged of the two provisions. The great re-promulgation litigation, from D. C. Wadhwa’s challenge to the Bihar practice onwards, arose under Article 213 and not under Article 123, and the reasons are visible in the design: State legislatures sit for fewer days, the temptation to govern by successive instruments is correspondingly stronger, and the proviso’s existence gives a challenge a second front to open.

Whose satisfaction it is: aid and advice and the parliamentary calendar

The satisfaction that Article 123 requires is in form the President’s. In substance, under the aid and advice scheme of Article 74, it is the satisfaction of the Union Council of Ministers. The President promulgates on ministerial advice, and the instrument is the government’s legislation in every practical sense: drafted in the government’s ministries, approved by the Cabinet, promulgated in the President’s name. Shubhankar Dam, whose study Presidential Legislation in India remains the standard scholarly account of the practice, documents how completely the device belongs to the government of the day: the formal satisfaction is a constitutional courtesy behind which the Cabinet legislates. The corresponding position under Article 213 follows from Article 163, with the Governor acting on the advice of the State Council of Ministers save where the Constitution requires discretion, and the ordinance power is not framed as a discretion. Any account that treats promulgation as a personal act of the President or the Governor mistakes the form for the substance, and any criticism of a particular ordinance addressed to the promulgating authority rather than to the government that advised it is addressed to the wrong desk.

The ordinary legislative process against which the power is defined is the passage of a Bill: introduction, consideration, passage by each House in the forms the Constitution prescribes, and assent, as the companion article on how a Bill becomes law in India sets out step by step. That process has two properties the ordinance route lacks: publicity in advance, since a Bill is debated in the open and usually examined in committee, and amendment by the legislature, since the text that emerges is the legislature’s text and not the drafting ministry’s. The ordinance route inverts both properties for its period of operation. The instrument takes effect in full on promulgation, and the legislature’s first sight of it as a working law comes when it is laid on reassembly, by which time the law has been operating, rights have been rearranged under it, and the replacing debate proceeds in the shadow of a functioning status quo that the instrument itself created. Defenders of the power reply that the inversion is the point: the circumstances the text contemplates are those in which waiting for the ordinary process would leave a gap in the law, and a gap in the law is also a public harm. The reply is at its strongest for measures whose content is time-critical and uncontentious, and at its weakest where the instrument’s content is a contested policy that the government expects the legislature to resist.

The calendar is where the arrangement’s deepest structural feature sits, and the feature deserves statement in the exact terms the series thesis requires. An ordinance may be promulgated only when the legislature is not in session. The summoning, adjournment, and prorogation of the legislature are acts performed on the advice of the government. The condition that makes the power available is therefore not an external fact that the government encounters; it is a state of affairs the government itself arranges. A government that wishes to legislate at once and without debate need not wait for a recess to occur. It can advise the prorogation of a sitting House, or simply allow a session to end and legislate in the gap it has itself created, and in the extreme form recorded in the public debate, Houses have been prorogued so that an instrument could issue without the embarrassment of a sitting legislature. The observation is structural and is made here without attributing motive to any particular government: the precondition for the power lies in the same hands as the power. The series has found the same configuration in other fields, a safeguard controlled by the party it constrains, and has recorded the same conclusion: such a safeguard does not bite. That is why, in practice, the restraints that operate on the ordinance power are not the recess condition but the restraints that operate after promulgation: the laying requirement, the six-week limit, the disapproval mechanism, and the prohibition on re-promulgation that the courts have built. The jurisprudence of the subject concentrates on what happens after promulgation for the same reason. There is little for a court to grip at the front end beyond the satisfaction itself, and the satisfaction, as the next sections show, is a narrow handle.

What is the difference between adjournment and prorogation when the ordinance power is measured?

The difference is the difference between a pause and an ending. Adjournment suspends a sitting to a later hour or day within the same session, and a House adjourned is a House still in session, so the recess condition of Article 123 is not met by a mere adjournment and no ordinance may be promulgated in an adjournment gap.

Prorogation terminates the session itself, and only when the session has ended by prorogation, or the House stands dissolved, is the legislature out of session in the constitutional sense and the power available. Governments that wish to create the condition therefore act on the session, not on the sitting.

The distinction also explains the eve-of-session and day-after-prorogation cases that recur in the criticism of the practice. An instrument promulgated on the day a session is prorogued satisfies the letter of the condition, since at the moment of promulgation the House is not in session. An instrument promulgated days before a notified session likewise satisfies the letter. The criticism in both cases is not that the text has been broken but that its purpose, legislation for the interval that cannot be avoided, has been stretched to cover an interval that existed only because the government chose its boundaries. The courts have not made the timing of a promulgation relative to a session a ground of invalidity by itself, and the satisfaction doctrine examined below explains why: the timing is part of the circumstances, and the circumstances are for the advised satisfaction unless the narrow grounds of challenge are made out. The check on eve-of-session promulgation is political and arithmetical rather than judicial. A government that promulgates on the eve of a session must still lay the instrument, still face the six weeks, and still carry a replacing Bill or lose the law, and it must do these things in a session it has just been seen to dodge. Whether that is check enough is the standing question of the subject, and it is a question about legislative practice rather than about the meaning of Article 123.

What an ordinance may contain, and what it may not

The starting proposition is coextensiveness. The power under Article 123 is coextensive with the legislative power of Parliament, and the power under Article 213 with the legislative power of the State legislature. The formulation runs through the case law: the Supreme Court in K. Nagaraj v State of Andhra Pradesh, decided in 1985, described the ordinance power as plenary within its field like the power of the legislature to pass laws, subject to no limitations except those to which the legislative power itself is subject; and T. Venkata Reddy, in the same year, proceeded on the same footing. An ordinance may therefore do anything an Act in the same field may do. It may amend an Act, repeal an Act, override a judgment’s statutory basis by changing the law the judgment applied, impose a tax, create offences and penalties, establish authorities, and operate retrospectively where an Act could operate retrospectively. The taxing power illustrates the width: in R. K. Garg v Union of India in 1981 the Court upheld the Special Bearer Bonds scheme carried by ordinance, rejecting the argument that taxation was too solemn a matter for the instrument, and A. K. Roy’s reasoning the following year placed the conclusion beyond argument by treating the ordinance as law for every constitutional purpose that speaks of law.

The limits are then exactly the limits of the corresponding legislature, and they are four in practice. The first is the distribution of competence. Parliament cannot legislate on a State List subject outside the situations the Constitution specifies, in a proclaimed emergency, on a Rajya Sabha resolution in the national interest, at the request of States, or to implement international obligations; an ordinance of the President is under the same disability, and Article 123(3) makes a provision beyond competence void. The State position mirrors it through Article 213(3), and the Governor’s instrument is further confined by the proviso’s three cases, which function as an additional content limit in the reserved fields. The second limit is Part III. An ordinance is law under Article 13(3), is void to the extent of any inconsistency with fundamental rights under Article 13(2), and may be challenged on that ground like any enactment; the temporary character of the instrument is no answer to a rights challenge, since a right abridged for seven months is abridged. The third limit is the rest of the Constitution outside Part III: an ordinance cannot override a constitutional provision, cannot alter the constitutional distribution of powers, and cannot do indirectly through a deeming fiction what a statute could not do. The fourth limit is the most absolute and the least often stated: an ordinance cannot amend the Constitution. Constitutional amendment is not legislation within the ordinary competence that the power borrows; it is the constituent process of Article 368, with its special majorities and, for the federal features, State ratification, and none of it can be compressed into an instrument promulgated in a recess. A government that needs a constitutional change must wait for Parliament and carry it there.

Two special applications of the limits recur in practice and deserve separate statement. The first is money. An ordinance may impose or vary a tax, as R. K. Garg establishes, and may authorise expenditure in the ways legislation does, but the appropriation of money out of the Consolidated Fund of India is governed by the procedure of Articles 114 and 266, which runs through a vote of the Lok Sabha and an Appropriation Act; the ordinance route does not displace that procedure, and governments have accordingly used ordinances to change tax law in a recess while leaving appropriation to the session. The careful statement is that the power reaches the substance of fiscal legislation but not the forms by which the Constitution commits public money. The second application is retrospectivity. Because the instrument is legislation, it may operate retrospectively in civil and fiscal matters as an Act may, and may validate past action by changing the legal basis on which the action was taken, subject to the bar in Article 20(1) on retrospective criminal liability, which binds an ordinance exactly as it binds an Act. Retrospective ordinances have been a standard tool where a tax demand or a regulatory scheme has failed in court on a curable defect, and their validity is tested by the same rules that govern retrospective Acts.

What can an ordinance do on the day it is promulgated that a Bill introduced in the next session cannot?

It can change the law at once. A Bill introduced when Parliament reassembles must pass both Houses and receive assent before it alters a single legal relation, and until then the old law governs. That difference in timing is the whole reason the power exists.

An ordinance alters the law from the moment of promulgation: offences are created or removed, taxes become payable or cease, authorities come into existence, and transactions arranged under the new position are valid transactions. That immediacy is the whole practical content of the power, and it is why governments reach for the instrument when a court judgment, a market event, or an administrative deadline will not wait for a session.

Immediacy has a price that the debate on the power sometimes prices and sometimes does not. The law made at once is law made without the legislature’s scrutiny, and the six weeks after reassembly are spent debating a replacing Bill against a background the instrument has already arranged. Appointments made under it are held, taxes collected under it are banked, licences granted under it are acted on, and the replacing debate is therefore not a debate about whether to change the law in the abstract but about whether to disturb arrangements already in being. The Supreme Court’s insistence, in the satisfaction cases, that the power is exceptional is the judicial form of the same observation. An instrument whose effects begin before its scrutiny begins is a derogation from the legislative process even when it is constitutionally impeccable, which is why the conditions on its use are not pedantry and why the record of each use, laid before the Houses, matters.

The courts on the power: satisfaction, character, and abuse

The case law on the power answers four questions in sequence. What is the character of the power? How far may a court examine the satisfaction on which it is exercised? What follows, for the product, from the instrument’s temporary life? And what happens when the power is used not for an interval but as a system? Each question has a leading decision, and the four decisions together give the doctrine this pillar owns, with the fourth question’s full development reserved to the companion article.

The character question was settled in A. K. Roy v Union of India, decided at the end of 1981 and reported in 1982. The challenge concerned the National Security Ordinance and the preventive detention scheme it carried, and the argument pressed on the Court was that ordinance-making is executive action and that its product is not law within the meaning of Article 21. The Court rejected the argument on converging grounds. Article 13(3) includes ordinances in the definition of law. Article 367(2) directs that references in the Constitution to Acts of Parliament be construed, unless the context requires otherwise, as including ordinances. Article 123(2) gives the instrument the same force and effect as an Act, and the article stands in a chapter headed Legislative Powers of the President. The conclusion followed that the President’s power to promulgate is legislative in character, that an ordinance is law for Articles 13 and 21, and that the only difference the Constitution draws between an ordinance and an Act is the difference of duration. Justice A. C. Gupta dissented on aspects of the case, but the majority’s characterisation has governed the subject since: an ordinance is legislation, tested as legislation is tested, and assailable on the grounds on which legislation is assailable, including incompetence and violation of fundamental rights.

The satisfaction question has a longer arc, because the constitutional text was itself altered in the middle of it. In R. C. Cooper v Union of India in 1970, the bank nationalisation case, the Court held that the President’s satisfaction under Article 123 was not wholly beyond judicial scrutiny: the promulgation could be challenged on the ground that the satisfaction was mala fide, that no satisfaction in law existed because the authority had acted on an extraneous or dishonest basis. The decision did not make the adequacy of the circumstances reviewable; it kept the door open at the threshold of good faith. Parliament’s response, in the Emergency period, was to close the door by text. The Constitution (Thirty-eighth Amendment) Act, 1975 inserted clause (4) into Article 123, and a corresponding clause into Article 213, providing that the satisfaction of the President or the Governor, as the case may be, shall be final and conclusive and shall not be questioned in any court on any ground. For the period that clause stood, the satisfaction was unreviewable as a matter of constitutional text, and the same amendment made parallel provisions for the emergency articles, a family of ouster clauses that belongs to the constitutional history of the year 1975. The Constitution (Forty-fourth Amendment) Act, 1978 omitted clause (4) from Article 123 and its State counterpart, the omission taking effect on 20 June 1979, and with the omission the position returned to the Cooper footing: satisfaction is a condition of the power, is in principle examinable on the narrow grounds Cooper identified, and is not examinable on the adequacy of the circumstances that produced it. A. K. Roy, decided after the omission, recorded that the challenge on mala fides remained available in principle, and the narrowness of the ground should not be mistaken for its absence. In practice, a challenger must plead and support an allegation that the satisfaction was not honestly formed or was formed for a purpose foreign to the article, and the record shows how rarely such a plea succeeds. But the doctrinal location of the door, shut by amendment in 1975 and reopened by amendment in 1978, is part of the power’s definition, and an account that omits the two amendments misstates the satisfaction doctrine in both directions.

What did the Supreme Court decide in D. C. Wadhwa’s case in 1987?

The Court held that the practice before it, the repeated re-promulgation of ordinances in Bihar over many years without placing them before the legislature for enactment, was a fraud on the Constitution and a subversion of the democratic legislative process.

The power to promulgate, the Court held, is an exceptional power to meet an extraordinary situation; it cannot be perverted to serve political ends, and it cannot be used as a substitute for the law-making power of the legislature. The one ordinance directly before the Court in the litigation was struck down, and the doctrine the case established governs the systematic use of the power rather than any single instrument.

The effects question belongs in its development to the companion article, but its 1985 station on the line belongs to this pillar, since the decision is also a principal authority on character. In T. Venkata Reddy v State of Andhra Pradesh, an Andhra Pradesh ordinance had abolished the posts of part-time village officers with effect from its commencement in January 1984; the ordinance was not replaced by an Act, and the abolished officers argued both that the instrument was invalid for non-application of mind and that, the instrument having ceased to operate, the posts must be treated as revived. The Court rejected both arguments. On character, the power to issue an ordinance is not an executive power but the power of the executive to legislate; a legislative act cannot be invalidated for non-application of mind or mala fides in the manner of an administrative act, since the legislature’s motives in passing a law are not examinable, and the ordinance stands in the legislature’s shoes. On effects, the Court read the words of the cessation clause exactly: the article says the instrument shall cease to operate, not that it shall be void from the beginning; it is therefore effective until it ceases, and a completed abolition effected while it operated was not undone by the cessation. The broad proposition that the consequences of an ordinance endure its expiry, as the consequences of a temporary Act were held to endure in the earlier line of authority, was substantially reconsidered by the seven-judge bench in Krishna Kumar Singh v State of Bihar in 2017, which held that the survival of rights and liabilities is not automatic and depends on the nature of the action taken and on considerations of public interest and constitutional principle; that reconsideration, and the re-promulgation doctrine it completed, are the property of the companion article and are stated here only to fix the 1985 decision’s place on the line.

The factual record of abuse: Bihar and the ordinance raj

The Wadhwa litigation is the case in which the factual record of the power’s abuse was placed before the Court in systematic form, and the record matters as much as the holding. Dr D. C. Wadhwa, a professor of economics whose research on land tenures in Bihar had taken him through the State’s legislative output, compiled the practice and published it before he petitioned. The figures he laid before the Court, and which the Court recorded, were these: between 1967 and 1981 the Governor of Bihar promulgated 256 ordinances; of these, 69 were re-promulgated repeatedly, with the prior permission of the President where the proviso required it; and individual instruments were kept alive by successive re-promulgation for periods running from one year to fourteen years, the legislature being presented with none of them for enactment in the ordinary way. The executive had, in substance, taken over the law-making function in the fields the instruments covered, and had done so not in defiance of the constitutional text at any single step, since each re-promulgation observed the forms, but in defiance of the scheme the text establishes. The majority, in a judgment led by Bhagwati CJ, held the practice to be a fraud on the Constitution: the ordinance power is conferred to meet an extraordinary situation, the re-promulgation of an instrument the legislature has not been asked to enact converts the exception into the system, and a colourable exercise of that kind is a subversion of the democratic legislative process. The Court struck down the one ordinance under direct challenge, the Bihar Intermediate Education Council Ordinance, and left the general doctrine to do its work prospectively. Three decades later, in Krishna Kumar Singh, a seven-judge bench converted Wadhwa’s condemnation into a holding of law on re-promulgation and completed the doctrine; that completion is the companion article’s subject, and this pillar takes from Wadhwa only what the power’s own account requires, which is the demonstration that the conditions and the clock, though real, do not by themselves prevent systematic misuse, and that the doctrine against misuse had to be built by the courts after the fact.

The pattern of use: frequency, sittings, and the worked examples

The frequency of the power at the Centre can be stated with the counting discipline the subject requires, attributed to a named source and a defined period. PRS Legislative Research, in its account of legislation in the Fifteenth Lok Sabha, records that 637 ordinances were promulgated at the Centre from the beginning of the first Lok Sabha in 1952 up to 2014, and its vital statistics for the Sixteenth Lok Sabha record 45 ordinances in that Lok Sabha’s term alone. The average across the long period is more than ten instruments a year, sustained across governments of every party and coalition shape, which is the fact that disciplines the rhetoric on both sides: the power is not the aberration of one administration but a standing instrument of Indian governance, used most heavily in periods of coalition instability and legislative congestion, and used in every period without exception. M. R. Madhavan, the co-founder of PRS Legislative Research, writing on the institution’s behalf, drew the reform conclusion from the same record: a facility conceived for rare use and exercised more than ten times a year has outgrown its conception, and his suggestion was that the facility might be repealed and Parliament summoned for urgent business instead. The suggestion has not been adopted, and the figure has not fallen to the level the conception implies, a pairing the reader should hold together: the record supports the criticism of frequency, and the frequency persists across the governments that voice the criticism when out of office.

The other half of the pattern is the calendar. The availability of the power expands as sittings contract, and sittings have contracted over the life of the Constitution. The first Lok Sabha sat for an average of 135 days a year; compilations of the parliamentary record put the average for the period from 1952 to 1970 at 121 days a year, and the average since the year 2000 at about 68 days a year. Each day not sat is a day on which the recess condition can hold, and each session not called is a gap in which the instrument can work. The contraction in sittings is a matter of institutional record, not of party: it has proceeded under every government, and it converts the ordinance from a bridge across an occasional gap into a parallel channel available for most of the year. A full account of the subject has to keep the two trends, frequency and sittings, in the same frame, since the constitutional question they jointly pose is whether a power designed for a legislature that sat for a third of the year retains its intended character when the legislature sits for a fifth of it.

Two Centre examples show the power working as designed and straining the design, and both are matters of public record. The Criminal Law (Amendment) Ordinance of 2013 was promulgated on 3 February 2013, in the interval after the Justice Verma Committee’s report on the law of sexual offences, and was replaced within weeks by the Criminal Law (Amendment) Act of 2013 when Parliament met: the instrument bridged a recess at a moment of demonstrated public urgency, and the replacing Act modified its terms in debate, which is the sequence the Constitution contemplates. The National Food Security Ordinance of 2013 was promulgated on 5 July 2013, days before the monsoon session, carrying a major welfare entitlement that had been pending as a Bill; it too was replaced by an Act in the session that followed, but the timing, hard against the session, drew the standing criticism that the recess had been used to pre-empt the debate rather than to bridge it. The two examples are a study in the limits of doctrinal analysis: the instruments are indistinguishable in form, both were replaced by Acts, and the constitutional objection attaches not to their validity but to the judgment about necessity that the satisfaction embodies and that the courts will not second-guess.

The example that belongs to this pillar’s cross-reference structure, because it shows a subject carried by successive ordinances over an extended period at the Centre, is the amendment of the land acquisition law. The Right to Fair Compensation and Transparency in Land Acquisition, Rehabilitation and Resettlement Act, 2013, which the series treats in its pillar on the LARR Act of 2013, was amended by an ordinance promulgated on 31 December 2014 that removed the consent and social impact assessment requirements for defined categories of acquisition. The replacing Bill passed the Lok Sabha and stalled in the Rajya Sabha; the ordinance was re-promulgated on 3 April 2015 in the form the Lok Sabha had approved, and re-promulgated again, as a second ordinance, on 30 May 2015. The sequence ended on 31 August 2015, when the government allowed the instrument to lapse rather than promulgate it a fourth time, and the 2013 Act continued unamended in the contested respects. The episode is the Centre’s clearest modern instance of the Wadhwa pattern, and its ending is as instructive as its course: what stopped the succession was not a court and not the clock, but the political cost of being seen to legislate by succession, which is the check the constitutional scheme in practice relies on when the legal checks run out.

The tribunal field supplies the contrasting example of the power used for a single bridge and then converted. The Tribunals Reforms Ordinance was promulgated on 4 April 2021, abolishing a set of appellate tribunals and transferring their work to the High Courts and the commercial courts, and it was replaced in the monsoon session by the Tribunals Reforms Act of 2021, deemed to operate from the ordinance’s date so that the legal position ran without a break. The instrument’s single promulgation, prompt replacement, and continuity of effect represent the orthodox use; the litigation that followed the Act, in which the Supreme Court struck down several of its provisions on tenure and appointment, belongs to the tribunal record and not to this pillar, but the ordinance phase of the episode is the model the text contemplates. The figures that frame these examples, the 637 instruments on the PRS count and the session records behind them, are conveniently assembled in tabular form for presentation, and a PowerPoint viewer serves readers who work through the session-by-session record in slide form.

The case for the power, and the case against it

The case for the power, at its strongest, is the case Ambedkar made in the Constituent Assembly, updated with the modern calendar. A legislature sits for a limited part of the year. Between its sessions, events do not pause: a court judgment may open a hole in a taxing statute that bleeds revenue daily; a financial institution may fail in a way that requires an immediate legal framework; an international obligation may fall due; a public order or public health measure may be needed in days. In each case the alternatives to the ordinance are worse than the ordinance. Continuous sittings are not available, since the members of a legislature in a large federal country have constituencies, committees, and the rest of the work of representation. A recall for every urgency would convert the exception into a standing arrangement indistinguishable from continuous sittings. And executive action without legal authority, the third alternative, is the one the Constitution exists to exclude: the power’s defenders are entitled to insist that a legal instrument promulgated in the open, laid before the Houses, and killed by the calendar unless the legislature adopts it, is not executive lawlessness but its opposite. The safeguards the defence relies on are real, and they are the post-promulgation safeguards: the laying requirement ensures the legislature confronts the instrument; the six weeks ensure it confronts it soon; the disapproval mechanism ensures it can kill it without the labour of repeal; and the bar on re-promulgation, as completed in 2017, ensures that what the legislature has not adopted cannot be renewed by the executive’s own act.

The case against, at its strongest, is the case the figures make when the two trends are held together. The power was designed for a legislature that sat for 135 days a year and an executive that legislated by exception; it operates in a system where the legislature sits for half that time and the exception issues more than ten times a year. The satisfaction, the only front-end condition, is controlled by the government that benefits from it and is reviewable only on the narrow ground of mala fides, which no challenger has the practical means to establish against a Cabinet’s collective assertion of necessity. The recess condition is in the government’s own gift, through its control of summoning and prorogation. The post-promulgation safeguards, on which the defence rests, are weaker than they look: the six weeks expire unless the opposition can carry a disapproval resolution in both Houses, a burden that falls on the critics rather than on the government; the replacing Bill is debated against facts the instrument has created; and the consequences of a lapsed instrument, in taxes collected, appointments made, and prosecutions launched, are in substantial part irreversible in practice whatever the doctrine says. Shubhankar Dam’s scholarly verdict, built on his count of the practice, is that the Indian executive legislates as a matter of routine and that the constitutional exception has become a third chamber in function if not in name. Madhavan’s reform suggestion, repeal of the facility in favour of summoning Parliament, is the institutional form of the same verdict. The Sarkaria Commission’s counsel of sparing and exceptional use at the State level is the moderate form of it.

The verdict this pillar records is the one the structure of the text supports. The power is constitutional in design and has been constitutionally necessary in identifiable instances, of which the criminal law amendment of February 2013 is the cleanest modern case; a constitution without the power would either leave gaps in the law during recesses or force the legislature into a sitting pattern no large democracy maintains. But the safeguards that carry the design are the safeguards after promulgation, and their strength depends on a legislature willing to use the six weeks as a genuine decision point rather than a formality, and on the political cost of succession, which is what ended the land acquisition sequence in 2015. Where the legislature treats the laying as a ritual and the replacing Bill as a foregone conclusion, the power operates as Dam describes, and no doctrine the courts have built can substitute for the legislative will the design presupposes. The text is not the weakness of the arrangement. The weakness is that the condition at the front is self-administered and the checks at the back are self-executing only against a legislature that chooses to work them.

The satisfaction in practice: what a challenger must show

The doctrine that the satisfaction is examinable on the ground of mala fides is easy to state and hard to use, and an account of the power that stops at the doctrine misleads the practitioner. A challenger does not obtain review by asserting that the circumstances did not warrant immediate action. The court will not weigh the circumstances, because the text commits the weighing to the advised satisfaction, and A. K. Roy’s settlement of the power’s character means that the ordinary administrative law grounds, non-application of mind, irrelevant considerations in the administrative sense, unreasonableness of the Wednesbury kind, do not transfer to a legislative act. What remains is a plea that the satisfaction was not the satisfaction the article requires at all: that it was formed dishonestly, for a purpose foreign to the article, or on a basis that shows the authority never directed itself to the existence of circumstances requiring immediate action. Such a plea must be pleaded with particulars and supported with material, and it is directed at a satisfaction formed collectively by a Cabinet whose deliberations are protected by Article 74(2), which bars a court from inquiring whether any, and if so what, advice was tendered by the Council of Ministers. The challenger must therefore build the case from the outside: from the timing of the promulgation against the parliamentary calendar, from the history of the subject in the legislature, from statements of the government’s own members, and from the relationship between the instrument’s content and any exigency that can be identified. The Wadhwa record shows the kind of material that moves a court, and it is material of a course of conduct over years, not material about a single instrument’s merits.

The practical consequence is a division of labour between the courts and the legislature that the text does not state but the doctrine produces. A single ordinance, however contestable its necessity, is in practice beyond judicial reach on the satisfaction ground unless the government’s own record betrays it. A pattern of ordinances, a succession on the same subject, a promulgation that follows a legislative rejection, is where doctrine gains purchase, because the pattern supplies the particulars the single case lacks and because the fraud-on-the-Constitution reasoning of Wadhwa and its completion in 2017 attach to the pattern rather than to the instrument. Governments have understood the division perfectly well. The discipline on a single promulgation is political, exercised through the laying, the debate on the replacing Bill, and the press of the session; the discipline on a pattern is legal, and it arrives late, after the pattern has done its work. A reader forming a judgment on any particular ordinance should therefore ask two questions in the right order: first, whether the instrument is within competence and consistent with Part III, which are questions a court can answer at once; and second, whether the promulgation belongs to a pattern that the fraud doctrine reaches, which is a question about the government’s conduct across sessions and not about the instrument in isolation.

R. C. Cooper in depth: the ordinance of July 1969

The satisfaction doctrine’s first station deserves a fuller statement, because the facts of R. C. Cooper v Union of India contain, in embryo, every later controversy. On 19 July 1969 the President promulgated the Banking Companies (Acquisition and Transfer of Undertakings) Ordinance, by which fourteen major commercial banks were nationalised. Parliament’s monsoon session was about to begin, and the instrument was promulgated on the eve of it; a replacing Act followed within the session in substantially the same terms. The challenge to the Act reached an eleven-judge bench, and on 10 February 1970 the Court, by a majority of ten to one, struck the Act down, holding that it violated the guarantee of compensation and the freedoms the Constitution then secured in Articles 19 and 31. The part of the case that belongs to this pillar is the treatment of the promulgation. The petitioners attacked the ordinance route itself, and the majority held that the President’s satisfaction under Article 123 was open to challenge where it was shown to be mala fide: the power is conditioned on a satisfaction honestly formed that circumstances require immediate action, and a satisfaction that is no satisfaction in law does not satisfy the condition. The Court did not strike down the measure on that ground; the Act fell on compensation and rights grounds. But the door Cooper opened, and which the Thirty-eighth Amendment then shut by text and the Forty-fourth reopened, is the door through which every later satisfaction argument has had to pass, and the eve-of-session timing of July 1969 remains the standard illustration that promulgation within days of a session is not, without more, a ground of invalidity: the Court examined the promulgation and proceeded to the merits, which is where the case was won and lost.

The case also fixes a point about the relationship between an ordinance and its replacing Act that recurs in practice. The ordinance and the Act were in substance one measure, and the challenge was fought on the Act because the Act was the operative law by the time the litigation matured. An ordinance that is promptly replaced transfers the constitutional contest to the replacing Act, and the promulgation’s own validity drops out of the case unless the succession itself is the complaint. That transfer explains a feature of the reports that puzzles first readers: the great ordinance cases are so often cases about Acts, about detention statutes, about abolished posts under successor instruments, because the life of a single instrument is shorter than the life of a constitutional challenge, and the litigation lands on whatever carries the measure when the court is ready to decide.

Article 213 in operation: instructions, reservation, and State practice

The proviso to Article 213(1) is usually recited and rarely traced in operation, and tracing it shows how the State power differs from the Union power in daily practice as well as in text. A State government’s legislative department, preparing an ordinance, must first run the proviso’s three questions against the content. Would a Bill in these terms have needed the President’s previous sanction for introduction? The class is defined by specific constitutional provisions, of which the restriction on State legislation affecting trade, commerce, and intercourse is the standard example: a Bill imposing reasonable restrictions in the public interest on the freedom of trade under Article 304(b) requires the previous sanction of the President before introduction. Would the Governor, at the assent stage, have deemed it necessary to reserve the Bill? That judgment engages the categories in which reservation is conventional, including Bills that derogate from the powers of the High Court, Bills on Concurrent List subjects that conflict with Union law, and Bills attracting the specific reservation provisions scattered through the Constitution. Would an Act in these terms have been invalid without reservation and assent? That is the repugnancy case in its pure form, where a State law on a concurrent subject conflicts with a Union law and can prevail in the State only through the Article 254(2) route of reservation and assent. Where any of the three answers is yes, the file must go to the Union for instructions before promulgation, and the ordinance waits on a decision in Delhi in which the State government is a supplicant. The delay is itself a discipline: the State power, in the reserved fields, cannot deliver the one thing the ordinance exists to deliver, which is speed.

The operation of the State power in the fields outside the proviso is disciplined by a different factor, the thinness of State legislative calendars. Many State assemblies sit for fewer days than Parliament, and some sit for only a few weeks in a year, divided into short sessions. The intervals between sessions are correspondingly long, and the temptation the Bihar record exposed, to carry subjects by successive instruments across those intervals, is a standing feature of State governance rather than a historical episode. The re-promulgation doctrine was built on State facts for that reason, and its application since 2017 has continued to be sought most often against State successions. A State government also faces a constraint the Union does not: its replacing Bill must pass a legislature that may sit too briefly to carry it within the six weeks, so the arithmetic of Article 213 presses harder on States with short sessions, and the pressure is one cause of the succession habit the doctrine condemns. The design assumes a legislature that returns in time and sits long enough to decide; where the assumption fails, the clock and the calendar between them manufacture the abuse, which is why the commissions on Centre-State relations, Sarkaria in 1988 and the Punchhi Commission reporting in 2010, addressed the ordinance power as part of the larger question of how State legislatures and Governors work, and counselled restraint in its use alongside reforms to sitting patterns and to the Governor’s office itself.

A worked anatomy: one ordinance through its life

The discipline of the subject is easiest to hold when it is followed through a single life, and the Criminal Law (Amendment) Ordinance of 2013 supplies a complete and public one. The Justice Verma Committee, appointed after the Delhi gang rape of December 2012, reported in January 2013 with recommendations for the reform of the law on sexual offences. Parliament was in recess; the budget session was weeks away. On 3 February 2013 the President, on the advice of the Union Council of Ministers, promulgated the ordinance, which rewrote the relevant provisions of the Indian Penal Code, the Code of Criminal Procedure, and the Indian Evidence Act with effect from that day. From 3 February 2013, therefore, the criminal law of the country stood changed: offences were defined in the ordinance’s terms, and investigations and prosecutions from that date proceeded under the amended law, because the instrument had the force of an Act from the moment of promulgation. The recess condition had been satisfied by the calendar; the satisfaction, that circumstances required immediate action, was the Cabinet’s response to a public emergency of confidence in the criminal law, and it is the kind of satisfaction the doctrine leaves alone.

Parliament reassembled for the budget session later in February, and the ordinance was laid before both Houses, as Article 123(2) requires. The six-week period began to run from the reassembly. The government introduced the replacing Bill, which was debated and modified: the enacted measure differed from the ordinance in particulars that the parliamentary process altered, which is the replacing process working as the design intends, the legislature legislating on a text the executive had put into interim force. The Criminal Law (Amendment) Act, 2013 received the President’s assent on 2 April 2013, within the six weeks, and was deemed to have come into force on 3 February 2013, the date of the ordinance, so that the legal position ran continuously from promulgation through enactment without a gap in which the old law revived. The ordinance itself was repealed by the Act as part of the standard replacing form. Every stage of the clock can be verified against the public record of that sequence: condition, promulgation, immediate operation, reassembly, laying, the six weeks, replacement with modification, continuity of effect. A reader who tests the step table in this article against the sequence will find each step where the table puts it, which is the sense in which the 2013 amendment is the power working exactly as the text designs it to work, and the contrast with the land acquisition sequence of 2014 and 2015, where the replacing process failed in one House and the instrument was renewed instead, is the contrast between the design and its strain.

Why the six weeks run from reassembly: the design, defended and doubted

The choice to run the six weeks from reassembly rather than from promulgation is the design decision on which the whole control structure rests, and it is worth understanding rather than merely memorising. A period running from promulgation would be a period that expired largely in the dark, during a recess, with the legislature unable to act on the instrument until it returned, possibly after the period had closed. By running the period from reassembly, the text guarantees the legislature a fixed working window, six weeks of its own sitting time, in which the instrument is before it, the disapproval resolution can be moved, and the replacing Bill can be carried. The Explanation’s later-date rule for staggered reassembly closes the one gap in the guarantee. The six-month ceiling on the inter-session gap then performs the complementary function of bounding how long the instrument can operate before the guaranteed window opens. The two periods are a matched pair: one forces the legislature back, the other gives it time when it returns, and the common error of running the six weeks from promulgation dissolves the pairing and should be corrected wherever it appears, since it converts a legislative window into a stopwatch the executive can wait out in a recess.

The doubt about the design is the doubt about the window’s use. Six weeks is a long time in a modern session’s arithmetic; sessions are shorter than they were, and a government that commands the session’s agenda can consume the window in other business and present the replacing Bill late, compressing scrutiny into the days before expiry concentrates minds. The disapproval mechanism, the design’s early stop, requires the opposition to win a vote in both Houses, which for a government with a working majority in the lower House and any presence in the upper is a demand for the government’s defeat rather than a scrutiny device. Critics therefore describe the window as a ratification period with a foreordained end, and the description is accurate for the ordinary case of a majority government and an uncontentious instrument. The design’s answer is that the window was never the only control: it operates together with the prohibition on succession, so that a government which lets the window close without enactment loses the law and cannot renew it by its own act, and with the political accounting that the laying and the debate make public. On the Bihar facts, what failed was not the window but the succession discipline, which did not exist as law until the courts built it; on the land acquisition facts, what worked was the politics. The six weeks are best understood as the design’s forcing device, whose function is to compel a legislative decision, and which performs that function even when the decision is a formality, because the formality is public and the record of it is permanent.

The ordinance beyond India: the device in comparative outline

A short comparison fixes what is distinctive in the Indian arrangement, stated at the level of settled constitutional text in each case. The device is an inheritance of colonial governance, and the constitutions that share the inheritance show the family resemblance. The Constitution of Pakistan, in Article 89, empowers the President to promulgate ordinances when the National Assembly is not in session and circumstances require immediate action, in language that follows the Indian model closely; the instrument there operates for a period of four months, extendable once for a further four months by resolution of the Assembly, a fixed term of the kind the Indian design declined in favour of the reassembly clock. The Constitution of Bangladesh, in Article 93, likewise empowers the President to make ordinances when Parliament stands dissolved or is not in session, with the instrument required to be laid before Parliament at its first meeting and ceasing thirty days thereafter unless approved, a shorter legislative window than the Indian six weeks. The constitutions of the Westminster family that lack the colonial inheritance lack the device: the United Kingdom knows no general power in the executive to legislate by instrument in a recess, and legislation there waits for Parliament or proceeds under specific statutory authorities of defined scope, so that the British executive confronted with an urgency in a recess recalls Parliament rather than legislating. The comparison sharpens the Indian design’s two distinctive choices: the power is general, extending to every subject within competence rather than to scheduled emergencies; and its control is located not in a fixed life but in the session, the laying, and the window. A reader assessing reform proposals for the Indian power will find that most of them are, in substance, proposals to move the Indian design toward one of the neighbouring models, by fixing a term, by shortening the window, or by scheduling the subjects, and that the choice among them is a choice about where the control on executive legislation should sit.

Reform proposals and their record

The reform literature on the power is long and its adoption record is empty, and stating both halves of that sentence is part of an accurate account. The proposals recur in a small number of forms. The first is abolition, Madhavan’s suggestion on behalf of no one but the argument itself, that the facility be repealed and Parliament summoned when urgency strikes; its strength is that it restores the legislature’s monopoly, and its weakness is the sitting pattern, since a Parliament that averages the sitting days the modern record shows would be recalled so often that the distinction between sessions and sittings would blur. The second is a fixed short life on the neighbouring models, replacing the reassembly clock with a term of weeks or months from promulgation; its strength is certainty, and its weakness is that it weakens the guarantee that the legislature will confront the instrument in its own time. The third is a bar on re-promulgation written into the text, converting the doctrine of 1987 and 2017 into a clause; its strength is that it would relieve the courts of policing a practice by litigation, and its weakness, after 2017, is that it would add little to the doctrine the courts have already built. The fourth is a sitting-days reform, a constitutional or conventional floor on the days Parliament and the assemblies sit each year, which attacks the condition rather than the power, shrinking the recesses in which the power lives; commissions and commentators have returned to it for decades, and it is the proposal whose adoption would change the pattern of use most while amending the ordinance articles not at all. The fifth is procedural: a requirement that the statement of circumstances accompany the laying in a form the legislature can debate, converting the satisfaction from a recital into a case to be answered. None of these has been enacted. The pattern of non-adoption is itself evidence about the power’s place in the system: every government that has criticised the device in opposition has used it in office at rates the PRS record displays, and a device that no incumbent proposes to surrender is, whatever the literature says, a settled part of the working constitution. The series’ verdict on the debate follows from that record. Reform of the ordinance power that depends on the executive surrendering it will not arrive; reform that works through the legislature’s own calendar and procedures, the fourth and fifth forms, works with the grain of the design, since the design already locates its real controls after promulgation and in the legislature’s hands.

Reading the instrument: form, publication, and commencement

An ordinance is a formal instrument and its form carries information a careful reader uses. It is promulgated in the name of the President or the Governor and published in the Gazette of India or the State Gazette, and publication is the act that gives it operation: the instrument speaks from its publication, and its recital records the two constitutional foundations in the text’s own language, that the legislature is not in session and that the promulgating authority is satisfied that circumstances exist which render it necessary to take immediate action. The recital is formulaic and proves nothing by itself, since it follows the article’s words in every case, but its presence marks the instrument’s claim to the power, and a challenge to the satisfaction must go behind a recital that asserts the condition in terms. The instrument is numbered in the series of ordinances for the year, is cited by that number and year in later law, and ordinarily commences at once on publication unless it appoints a different day or is given retrospective operation, which it may be for the reasons the content treatment states. In structure it mimics an Act: short title, commencement clause, definitions where needed, operative provisions amending or enacting, and repeal and savings where it displaces existing law. The resemblance is deliberate, because the instrument must slot into the statute book it amends and be administered by authorities who work from texts of that shape.

The documentation that surrounds the instrument in the legislative process is the other half of its form. When Parliament reassembles, the ordinance is laid before both Houses, and the practice of the Houses requires the government to lay with it a statement explaining the circumstances that necessitated immediate legislation by ordinance, so that the satisfaction, unexaminable in its adequacy by a court, is rendered in narrative form to the legislature whose scrutiny is the design’s real control. The replacing Bill that follows is introduced with its own statement of objects and reasons, which conventionally records that the Bill replaces the ordinance, and the parliamentary record of the replacement, the debates, the amendments moved and carried, is the public account against which the promulgation’s necessity is judged in the political forum. A researcher tracing any particular instrument therefore works from three documents in sequence: the Gazette text of the ordinance, the explanatory statement laid with it, and the replacing Bill with its parliamentary history, and the divergences among the three are often the whole story of the measure, since content dropped between the ordinance and the Act marks the points on which the legislature’s scrutiny bit, and content that survives unchanged marks the measure the executive would have enacted without it.

Commencement questions under ordinances are a practical trap worth naming. Because an ordinance takes effect on promulgation unless it provides otherwise, acts done on the day of publication are done under the new law; because a replacing Act is ordinarily deemed to have come into force on the date of the ordinance, the continuity of the regime across the replacement is seamless in the standard case; and because an instrument that lapses without replacement leaves the prior law to operate from cessation, transactions in the window between cessation and any later legislation are governed by the restored position. The three positions are mechanical once stated, and the errors in practice come from assuming that the seamlessness of the standard case extends to the lapsed case, which is exactly the assumption the effects doctrine, in its developed form in the companion article, exists to test.

The federal distribution worked through examples

The competence limit rewards working through the distribution, since it is in application that Article 123(3) and Article 213(3) earn their place. A Union ordinance may legislate on any Union List subject and on any Concurrent List subject, as Parliament may. On a State List subject it may legislate only when Parliament itself could: during a proclaimed emergency, when the Rajya Sabha has resolved in the national interest, when two or more States have requested the legislation, or to give effect to an international agreement. Outside those gates, a provision in a Union ordinance on a State List subject is void under Article 123(3), and the voidness is of the provision, so that a substantially valid instrument can lose a clause to the distribution while standing as a whole. The mirror image at the State level is more frequently litigated in practice, because State instruments press against Union occupation of concurrent fields: a Governor’s ordinance on a Concurrent List subject that conflicts with a Union law falls into the repugnancy analysis, and the proviso’s third case is the bridge to the only cure, reservation and assent, which the ordinance route cannot supply for itself without instructions. The practical advice embedded in the text is that a State ordinance in a concurrent field occupied by Union legislation is the most fragile instrument in the whole scheme, exposed on competence, on repugnancy, and on the proviso at once.

A State worked example shows the machine in ordinary operation. In November 2020 the Governor of Uttar Pradesh promulgated the Uttar Pradesh Prohibition of Unlawful Conversion of Religion Ordinance, creating offences and a regulatory scheme around religious conversion, on the advice of the State government and while the State legislature was not in session. The subject drew on the State’s competence over public order and allied State and concurrent fields; no instructions question of the proviso’s kind was raised in the public record of the promulgation; and the instrument was replaced by an Act of the State legislature early in 2021, within the Article 213 sequence, so that the regime continued as enacted law. The example is unremarkable in its mechanics, and that is its value: it shows the State power performing its designed function of bridging to a session for a measure the government regarded as urgent, and it shows the content limits doing silent work, since the measure’s vulnerability, debated in the public argument that followed, lay in the fundamental rights analysis under Part III, where an ordinance stands in no better position than the Act that replaced it. The same template, promulgation, operation, laying, replacement, describes the great majority of instruments at both levels in ordinary times; the celebrated cases are the exceptions in which one element of the template, most often replacement, failed.

Failure modes: how an ordinance ends early, or never lives

The four outcomes in the step table are not equally frequent, and the distribution among them is part of the power’s real description. Replacement by an Act is the overwhelming norm for Union instruments: the government that promulgates ordinarily commands the lower House and carries the replacing Bill within the six weeks, and the instrument’s constitutional life ends in the Act’s continuity. Expiry without replacement is the failure the design prices in, and its modern exemplar is the land acquisition sequence’s ending in August 2015, where expiry was chosen by the government as the least costly exit from a succession it had decided not to continue. Withdrawal before the legislature decides is the quiet outcome, used where a government has misjudged a measure’s reception or where the exigency that prompted the instrument has passed; it attracts little attention precisely because it concedes the argument, and its existence in the text, an executive early stop exercisable at any time, is the answer to the suggestion that a promulgated instrument escapes its author’s control. Disapproval is the rarest outcome of the four. A resolution of disapproval requires the critics of the measure to carry both Houses within the six weeks, which in the ordinary arithmetic of a majority government means defeating the government on the floor; disapproval resolutions are moved and debated as part of the replacing process’s politics, but their carriage against a government that retains its majority is not the design’s working assumption. The working control is the debate itself and the deadline, and the honest account of the design states plainly that the disapproval mechanism functions less as a kill switch than as the occasion on which the government must defend the promulgation in public and on the record.

An ordinance may also never live in law at all, in the distinct sense of being stillborn on competence or rights grounds. A provision beyond the legislature’s competence is void from the start under the third clause of each article, and an instrument inconsistent with fundamental rights is void under Article 13 to the extent of the inconsistency; either defect may be declared by a court in a challenge brought while the instrument operates or, where consequences persist, after it has ceased. The practical difficulty is temporal: constitutional litigation moves more slowly than the seven and a half months of a maximum life, so a challenge to a single instrument commonly finds its target replaced or expired by the time of decision, and the court’s pronouncement then operates on the successor Act or on the surviving consequences. The doctrine has grown comfortable with that difficulty, deciding the validity questions as they bear on the live measure and on accrued consequences; but a challenger whose object is the promulgation itself, rather than the measure’s content or its residue, faces the arithmetic the whole subject is built on, that the instrument is designed to be gone before it can be judged, and that the system’s answer to that design is to judge patterns, successors, and consequences instead.

The replacing process in the Houses

The six weeks after reassembly have their own procedural life, which the constitutional text assumes and the rules of the Houses supply. The government lays the ordinance on the table of each House, with the statement of circumstances; a member who objects to the measure may give notice of a statutory resolution disapproving it, and the practice of the Houses is to take the disapproval resolution and the government’s replacing Bill together, so that the debate is a single contest over the measure’s merits in which the resolution voices the objection to the route and the Bill carries the question of the content. If the resolution is negatived, as it ordinarily is on the government’s numbers, the Bill proceeds through its stages; if the Bill is amended in passage, the Act that emerges supersedes the instrument in the amended terms, and the divergences between the ordinance and the Act are the legislature’s amendments made visible. Where the two Houses divide, the ordinary mechanisms of the legislative process apply to the replacing Bill, including the joint sitting for a deadlocked ordinary Bill; the ordinance route confers no advantage in that machinery, which is a point governments discover when a measure passes the Lok Sabha and stalls in the Rajya Sabha, as the land acquisition amendment did: the instrument can bridge to the session, but it cannot carry the session, and at the end of the six weeks the bridge ends whether or not the far bank has been reached.

Succession at the Centre, where the replacing process fails in one House, is illustrated by the securities legislation of 2013 and 2014. The Securities Laws (Amendment) Ordinance was promulgated in 2013 to enlarge the Securities and Exchange Board of India’s enforcement powers, and, the replacing Bill not having completed its passage, the instrument was re-promulgated, and re-promulgated again, before the replacing Act was enacted in 2014; PRS Legislative Research, in recording the Fifteenth Lok Sabha’s legislative account, noted the re-promulgations among the features of that Parliament’s working. The episode predates the 2017 judgment, and it is the kind of Centre practice the judgment’s doctrine governs: a succession occasioned by a Bill’s pendency in the legislature is distinguishable in its facts from the Bihar pattern of instruments never laid for enactment at all, but the distinction is one the government asserting it must be able to state, since the fraud doctrine fastens on the legislature’s exclusion from the decision, and a succession maintained while a Bill waits undecided in a House is a succession the House has not been allowed to complete. The prudent reading of the modern law, for a government and for a challenger alike, is that pendency explains a single renewal at most and explains nothing once the pattern becomes the method.

Counting ordinances honestly

The frequency debate is conducted with numbers, and the numbers need the same care as the text, because counting ordinances is not a neutral act. A succession of five instruments on one subject counts as five in the raw total and as one measure in substance; a count by instrument therefore overstates the number of distinct occasions on which the power was used, while a count by subject understates the number of times the legislature’s session was bypassed. The PRS figure of 637 instruments at the Centre from 1952 to 2014 is a count of instruments, honestly labelled, and the average of more than ten a year that it yields is an average of promulgation events, which is the right measure for the constitutional question, since each promulgation is a separate decision to legislate without the legislature. Dam’s scholarly count in Presidential Legislation in India proceeds on the same basis and reaches totals of the same order, and his analysis adds the distribution the raw total hides: the rate was not uniform, rising in the decades of unstable majorities and falling when single-party governments with full calendars legislated through Parliament in the ordinary way. The honest use of the figures in argument therefore requires three disciplines, all of which this article has observed: name the source, name the period, and say what is being counted. A figure without those three attendants is rhetoric, and the debate on the power has suffered from unattributed figures quoted across decades in which the underlying practice had changed.

The State figures require the same discipline and a further caution. Bihar’s 256 instruments between 1967 and 1981, with 69 kept alive by re-promulgation for periods of up to fourteen years, is a figure from the petitioner’s research recorded by the Supreme Court, and it is a figure about succession as much as about frequency: the constitutional offence in Bihar was not that the Governor promulgated often but that promulgation replaced legislation. State counts in the ordinary run are lower in drama and similar in structure: instruments cluster around the gaps in short legislative calendars, and the proportion that are replaced at the next session is the figure that separates a working bridge from a parallel legislature. No single all-States count commands the field the way the PRS count does for the Centre, and an account of the subject does better to say so than to manufacture one.

An ordinance is not delegated legislation

One confusion should be cleared before the boundary of the pillar, because it recurs in commentary and in examination answers alike. An ordinance is not a species of delegated or subordinate legislation. Rules, regulations, notifications, and bye-laws are made by authorities to whom a parent Act delegates power, and they are valid only so far as they stay within the delegation, a doctrine of ultra vires that polices them strictly. An ordinance has no parent Act. Its authority is the Constitution itself, its maker is the President or the Governor exercising a power the Constitution classifies as legislative, and the test of its validity is not fidelity to a delegation but the same test an Act must pass: competence, consistency with Part III, and consistency with the Constitution at large. The confusion arises because both instruments are made by the executive branch in personnel terms; but the executive wears two hats in the constitutional scheme, the administrative hat under which it makes subordinate legislation within Acts, and the legislative hat under which, in a recess and on satisfaction, it makes law directly under Articles 123 and 213. The remedies differ accordingly. Subordinate legislation is struck down for exceeding its parent; an ordinance is struck down, if at all, as an Act is struck down; and the satisfaction condition, which has no counterpart in the law of delegation, is the distinctive control the constitutional route carries in place of the parent Act’s limits.

Edge cases: dissolution, President’s Rule, and the courts’ interim hand

Three situations test the framework at its edges, and each is answered by the text once the framework is held firmly. The first is dissolution. The condition of Article 123 is that both Houses are not in session, and a dissolved Lok Sabha is not in session in any sense; the Rajya Sabha, a continuing body, may or may not be sitting. An ordinance may therefore be promulgated during a dissolution, when the Council of Ministers continues in office in its caretaker capacity until the new House is constituted. The instrument then takes its course under the new Parliament: it is laid before the Houses of the newly constituted Parliament when they assemble, and the six weeks run from that reassembly, so the legislature that judges the instrument is not the legislature whose recess made it possible. The arrangement is a faithful working of the text, and it sharpens the accountability question the subject keeps posing, since the government that advised the promulgation may not be the government that defends the replacement; the design’s answer is that the instrument belongs to the office and not to the administration, and that the new House’s six weeks are a real decision point at which a successor government must adopt, amend, or abandon its predecessor’s bridge. Caretaker conventions counsel restraint in the use of the power during a dissolution, since major policy by ordinance in the interval between a House’s death and its successor’s birth sits uneasily with the principle that the outgoing government’s mandate is spent; but the text imposes no caretaker exception, and the restraint is conventional rather than legal.

The second situation is President’s Rule in a State. Where a proclamation under Article 356 is in operation and the powers of the State legislature are exercisable by or under the authority of Parliament, State legislation proceeds under Article 357, which empowers Parliament to confer on the President the power of the State legislature to make laws; laws so made have the special endurance Article 357 prescribes, continuing in force until altered or repealed by the competent legislature. In that setting the Article 213 route is not the vehicle: the Governor’s ordinance power presupposes a State legislature in recess whose competence the instrument borrows, and where the legislature’s powers stand transferred by proclamation, the borrowing runs from Parliament’s authority instead. The distinction matters in practice because the products differ in their fate: a law made under Article 357 does not face the six-week clock of Article 213, and the confusion of the two routes in commentary produces errors about the duration of President’s Rule legislation that the text, read carefully, does not license.

The third situation is the courtroom during the instrument’s short life. A challenger who seeks to stop an ordinance in its operation asks the court for interim relief against a law in force, and the courts apply to that request the discipline they apply to legislation generally: there is a presumption of constitutionality, and interim stay of a statute’s operation is exceptional, granted where the court is persuaded that the balance of irreversible harm requires the law to wait for its judgment. Against an instrument whose whole life is measured in months, the discipline produces the temporal difficulty already noted: interim relief is sparingly given, final hearing takes longer than the instrument, and the contest migrates to the successor Act or to the consequences. The exceptions prove the structure. Where the challenge is to a succession, the court’s eye is on the course of conduct and the relief can reach the pattern; where the challenge is to content that a successor Act has carried forward, the litigation loses nothing by the transfer. What the challenger cannot ordinarily obtain is the one thing the design withholds from everyone, a preview of the satisfaction: the court will not suspend the instrument while it decides whether the Cabinet’s assertion of necessity was honest, on an application that must be built from outside the protected deliberations.

The two articles side by side: a synthesis of the power

The pillar’s account can be closed by holding the two provisions against each other in a single view, since that is the comparison the subject’s students and practitioners are finally asked to make. The grants are parallel: a satisfaction-conditioned power to legislate in a recess, with plenary force, a laying obligation, a six-week window from reassembly, and early endings by disapproval and withdrawal. The differences are four. The State power carries the proviso’s three instructions cases, and the Union power carries none. The State disapproval is perfected by the Assembly’s resolution agreed to by the Council where one exists, while the Union disapproval requires resolutions of both Houses as equals in the mechanism. The Union power operates against a Parliament whose inter-session gap is capped at six months by Article 85, and the State power against legislatures capped by Article 174 in the same terms but sitting, in many States, for far fewer days within the cap. And the Union power’s satisfaction was the subject of the Thirty-eighth and Forty-fourth Amendments in terms that applied to both articles together, so that the review doctrine is common to the two while the practice, the litigation, and the abuse record have concentrated in the States. Everything else that distinguishes the two powers in the world, the frequency of succession, the thinness of calendars, the friction of the Governor’s double posture, is a consequence of those four textual and institutional facts, and an analysis that keeps them in view will not go far wrong on any question the subject presents.

The wider family: Union Territories and Article 239B

The two articles at the pillar’s centre do not exhaust the Constitution’s ordinance powers, and the remaining member of the family confirms the design’s logic. For Union Territories, legislation is Parliament’s business in the first instance, and the President’s power under Article 123 reaches every subject on which Parliament may legislate for such a territory, since the competence the article borrows is Parliament’s own. Where a Union Territory has a legislature of its own, the recess between its sessions would leave the same gap the State articles fill, and for Puducherry the Constitution fills it expressly: Article 239B empowers the Administrator, when the Legislative Assembly of the Union Territory is not in session and circumstances require immediate action, to promulgate ordinances with the force of an Act of that Assembly, subject to the familiar apparatus of laying, cessation six weeks from reassembly, disapproval, and withdrawal. The borrowing is from the State model, not the Union one: Article 239B reproduces the discipline of presidential instructions, barring the Administrator from promulgating without instructions from the President in cases corresponding to the sanction, reservation, and invalidity cases of the Article 213 proviso, a pattern that reflects the territory’s position below the Union in the constitutional hierarchy, as the Governor’s proviso reflects the State’s position within the federal scheme. It was this family of provisions, the President’s, the Governors’, and the administrators’, whose satisfactions the Thirty-eighth Amendment made final together and the Forty-fourth Amendment released together, a reminder that the ordinance device in the Indian Constitution is not an isolated grant to the Union executive but a general technique of the scheme, deployed at each level of government with a supervision pattern fitted to that level’s place in the whole. Delhi, whose legislative arrangements rest on Article 239AA and the Government of National Capital Territory of Delhi Act, has no corresponding ordinance power for its Lieutenant Governor, and legislation for the capital in a recess of Parliament proceeds, where it must, under the President’s general power; the asymmetry is a deliberate consequence of the capital’s distinct constitutional settlement, and it completes the map of where in India law may be made by instrument and by whom.

The One Test with which this article began can be marked. An ordinance is legislation made by the executive when the legislature is not sitting: by the President under Article 123 on the satisfaction that circumstances require immediate action, by a Governor under Article 213 on the same satisfaction within the State field. Its conditions are the recess and the satisfaction; its clock is the six months that bound the recess and the six weeks that run from reassembly; its endings are replacement, expiry, disapproval, and withdrawal; its content is bounded by the competence of the legislature whose place it takes, by the fundamental rights, and by the bar on constitutional amendment. And the Governor’s power is narrower than the President’s in the three respects the proviso enumerates, each of which subordinates the State instrument to the President’s instructions where the ordinary legislative process would have engaged the President: where a Bill in the same terms would have required his previous sanction for introduction; where the Governor would have deemed it necessary to reserve such a Bill for his consideration; and where an Act in the same terms would have been invalid unless reserved for his consideration and assented to by him. A reader who holds those propositions, and the structural observation that the recess condition lies in the promulgating government’s own hands, holds the power as the Constitution grants it, as the courts have glossed it between 1970 and 1987, and as the record shows it used.

One further edge: the summoned but unsat House

The concept of a session does one more piece of work at the margin. A House is in session across the period that begins with its first sitting after being summoned and ends with prorogation or, for the Lok Sabha, dissolution; adjournments within the period do not break the session, as the adjournment and prorogation treatment in this article states. It follows that in the interval between the issue of summons for a session and the first sitting under it, the House is not yet in session, and the recess condition can still be satisfied: a promulgation in that interval is within the letter of Article 123. The interval is usually short, and the promulgation it permits belongs to the family of eve-of-session cases whose constitutional objection is one of purpose rather than of text. The distinction the courts’ doctrine draws serves here as well: the timing may evidence the character of a course of conduct when a pattern is challenged, but it does not by itself invalidate the instrument, and a reader testing a particular promulgation should record the interval accurately and argue from the pattern, not from the calendar alone.

For the practitioner advising on a live instrument, the framework resolves into a checklist that follows the article’s order. Confirm the recess from the Gazette and the session records of the relevant legislature, remembering that one House sitting does not close the Union power and that an adjournment is not a recess. Read the instrument against the distribution of legislative fields, provision by provision, since the third clause voids to the extent of incompetence and no further. Test the content against Part III as an Act would be tested, and, at the State level, run the proviso’s three questions before anything else, because an instrument that needed instructions and lacked them is exposed at its root. Fix the dates: the promulgation, the reassembly of each House, the later date where they differ, and the day the six weeks expire, since the instrument’s life and the replacing Bill’s deadline both hang on that arithmetic. Identify the outcome the instrument is travelling toward and the client’s position under each of the other three, with the effects question, where it matters, taken to the doctrine of the 2017 judgment through the companion article. An adviser who works the checklist in that order will find that the framework, for all the controversy of the power it describes, answers the ordinary questions completely, and reserves its difficulty for the single question the text itself leaves open, which is what survives the instrument when the instrument is gone.

The effects question, and the boundary of this pillar

Every route out of an ordinance except replacement by an Act raises the same question in a different temper. When the instrument ceases to operate, whether by expiry of the six weeks, by disapproval, or by withdrawal, what is the position of the things done under it while it operated? Taxes assessed and collected, appointments made and served under, contracts concluded, offences created and prosecuted, posts abolished and their holders sent home: the cessation clause says the instrument ceases to operate, and it does not say that the world is restored to its position on the day before promulgation. The doctrine’s journey runs from the early holding that the consequences of a temporary enactment endure its expiry, through Venkata Reddy’s reading of the cessation words in 1985, to the seven-judge reconsideration in Krishna Kumar Singh v State of Bihar in 2017, where the Court held that re-promulgation is a fraud on the Constitution, that the satisfaction is open to scrutiny, that the obligation to lay the instrument before the legislature is mandatory and its breach examinable, and that the endurance of acts done under a ceased instrument is not automatic but turns on the nature of the action and on public interest. That judgment is the leading authority on the modern law of the subject, and the companion article on the ordinance re-promulgation judgment states and analyses it in full: the questions whether a lapsed instrument’s consequences survive, whether an instrument never laid before the House can claim effect, and whether a succession of instruments can be attacked as a single course of conduct are its questions, and this pillar does not take them, beyond recording the headline propositions that fix the boundary.

Do acts done under an ordinance unravel when it ceases to operate?

Not automatically, and not never. The cessation of an ordinance ends the instrument’s operation from that point; it does not rewrite the period during which the instrument was law. That distinction, between ending a law and erasing its past, is the foundation of the whole effects doctrine.

Whether a particular act done in that period, an appointment, a collection, an abolition, continues to have legal effect after cessation is a further question, on which the older doctrine of enduring effects was substantially reconsidered in 2017: survival is not a foregone conclusion but depends on what was done, whether it was completed and closed, and whether public interest and constitutional principle require it to stand. The developed analysis belongs to the companion article on the leading judgment.

For the practitioner, the boundary of this pillar is the practical line. The power, its conditions, its clock, its content limits, and its character as legislation are stated here and may be relied on as the framework. The consequences of a particular instrument’s cessation in a particular matter, the prospects of a challenge to a succession of instruments, and the effect of a failure to lay are questions on which the 2017 judgment is the controlling authority and the companion article the controlling account in this series, and advice in a live matter should be taken on the facts with that judgment in hand. This article is a statement of the law’s framework, offered as legal information; it is not advice on any person’s case.

Frequently Asked Questions

Q: On what condition may an ordinance be promulgated?

Two conditions must hold together under Article 123. The first is a recess: the power is available at any time except when both Houses of Parliament are in session, so at least one House must not be sitting. The second is satisfaction: the President must be satisfied that circumstances exist which render it necessary for him to take immediate action. No emergency, war, or named crisis is required, and the word emergency does not appear in the article. The satisfaction is in substance that of the Council of Ministers, on whose advice the President acts. Article 213 imposes the same two conditions for a Governor in relation to a State legislature.

Q: Is an ordinance equivalent to an Act of Parliament?

In force and effect, yes; in origin and duration, no. Article 123(2) provides that an ordinance shall have the same force and effect as an Act of Parliament, and the Supreme Court in A. K. Roy v Union of India held that the power is legislative in character and the product is law for the purposes of Articles 13 and 21, subject to the same constitutional limitations as an Act. The differences are that an ordinance is made by the executive on ministerial advice rather than enacted by the legislature after debate, and that its life is bounded by the six weeks from reassembly unless it is replaced by an Act, while an Act endures until repealed.

Q: How long does an ordinance last?

An ordinance operates from its promulgation until the earliest of four events: replacement by an Act of Parliament, withdrawal by the President, disapproval by resolutions of both Houses, or the expiration of six weeks from the reassembly of Parliament. The six weeks run from reassembly and not from promulgation, which is the point most accounts get wrong. Where the two Houses are summoned to reassemble on different dates, the Explanation to Article 123(2) directs that the six weeks be reckoned from the later of those dates. Since six months is the longest gap the Constitution permits between sessions, the outer limit of an ordinance’s life is six months plus six weeks.

Q: What must happen to an ordinance when the House reassembles?

The ordinance must be laid before both Houses of Parliament, as Article 123(2) requires, and the government lays with it a statement explaining the circumstances that made immediate legislation necessary. From the reassembly, the six-week period runs. Within that window three things may happen: the Houses may pass resolutions disapproving the ordinance, in which case it ceases to operate on the passing of the second resolution; Parliament may enact a replacing Bill, in which case the law continues as an Act; or neither may occur, in which case the ordinance ceases to operate of its own force at the end of the six weeks. The laying is a constitutional obligation, not a courtesy.

Q: How does a Governor’s ordinance power differ?

The Governor’s power under Article 213 mirrors the President’s in its conditions, its force, its laying requirement, and its six-week clock, but it is narrower in one structural respect: the Governor cannot promulgate without instructions from the President in three cases. The cases are where a Bill containing the same provisions would have required the President’s previous sanction for its introduction into the State legislature; where the Governor would have deemed it necessary to reserve such a Bill for the President’s consideration; and where an Act containing the same provisions would have been invalid unless reserved for the President and assented to by him. Article 123 has no counterpart restriction.

Q: Can the satisfaction behind an ordinance be questioned in court?

Yes, on narrow grounds. In R. C. Cooper v Union of India in 1970 the Supreme Court held that the President’s satisfaction under Article 123 was not wholly beyond judicial scrutiny and could be challenged where it was mala fide. The Constitution (Thirty-eighth Amendment) Act, 1975 then inserted a clause making the satisfaction final and conclusive and not questionable in any court, and the Constitution (Forty-fourth Amendment) Act, 1978 removed that clause, with effect from 20 June 1979, restoring the Cooper position. A court will not examine whether the circumstances were adequate; the challenge must show that the satisfaction was dishonestly formed or formed for a purpose foreign to the article, which is a demanding plea in practice.

Q: How frequently have ordinances been promulgated?

The record is one of routine use across every period of the Constitution’s life. PRS Legislative Research records that 637 ordinances were promulgated at the Centre between the first Lok Sabha in 1952 and the year 2014, an average of more than ten a year, and its vital statistics for the Sixteenth Lok Sabha record 45 ordinances in that single Lok Sabha’s term. The rate has varied with the stability of governments and the congestion of the legislative calendar, rising in periods of coalition instability. In the States the outstanding record is Bihar’s, where the Supreme Court in D. C. Wadhwa’s case recorded 256 ordinances between 1967 and 1981, many kept alive by re-promulgation.

Q: What are the limits on what an ordinance can do?

The power is coextensive with the legislative power of the corresponding legislature, and its limits are that legislature’s limits. An ordinance cannot contain a provision the legislature itself could not enact: Article 123(3) makes void any provision Parliament would not be competent to enact, and Article 213(3) does the same at the State level. It is subject to the fundamental rights as law under Article 13, and may be struck down for violating Part III. It cannot amend the Constitution, since amendment is the constituent process of Article 368 and not ordinary legislation. Within those limits it may do anything an Act may do, including amending or repealing Acts and imposing taxes.

Q: Who actually decides that an ordinance will be promulgated?

The Council of Ministers decides, and the President promulgates. Under the aid and advice scheme of Article 74, the President acts on the advice of the Union Council of Ministers headed by the Prime Minister, and the satisfaction Article 123 names is in substance the Cabinet’s satisfaction that immediate action is necessary. The instrument is drafted in the government’s ministries, approved by the Cabinet, and issued in the President’s name. The position is the same in a State, where the Governor acts on the advice of the State Council of Ministers under Article 163. Criticism or defence of a particular ordinance is therefore properly addressed to the government of the day, not to the promulgating authority personally.

Q: Can an ordinance be promulgated when one House of Parliament is in session?

Yes. Article 123(1) bars promulgation only when both Houses of Parliament are in session. The text reads “except when both Houses of Parliament are in session”, so if either House is not sitting, the condition is satisfied and the power is available. In practice the two Houses are ordinarily summoned and prorogued together, so the situation seldom arises; but where the Houses are recalled on different dates, an ordinance may lawfully issue while one House is already sitting. The Explanation to Article 123(2) protects the instrument in that situation by reckoning the six weeks from the later of the two reassembly dates.

Q: What is the longest an ordinance can stay in force?

Seven and a half months at the outside, on the Constitution’s own arithmetic. Article 85 provides that six months shall not intervene between the last sitting of a House in one session and the first sitting in the next, so the recess in which an ordinance operates cannot lawfully exceed six months. Once Parliament reassembles, Article 123(2) allows the instrument a further six weeks. An ordinance promulgated at the start of the longest lawful recess and left to run its full course therefore lives for six months plus six weeks, the maximum the Supreme Court stated in A. K. Roy. Disapproval or withdrawal can always end it sooner, and a replacing Act converts rather than extends it.

Q: Can an ordinance amend or repeal an Act of Parliament?

Yes. The ordinance power is coextensive with the legislative power of Parliament, and an instrument under Article 123 may amend, repeal, or supplement any Act within Parliament’s competence, with the same effect as an amending or repealing Act for so long as the ordinance operates. Most ordinances in practice are amending instruments of exactly this kind, altering an existing statute in a recess ahead of a replacing Bill. If the replacing Act is enacted, the amendments continue in statutory form; if the ordinance ceases to operate without replacement, the amendments fall with it and the earlier position revives, subject to the doctrine on acts done while the instrument operated, which the companion article on the re-promulgation judgment develops.

Q: Can an ordinance amend the Constitution?

No. A constitutional amendment is not ordinary legislation. It is made under the constituent power in Article 368, which requires a special majority in each House of Parliament and, for the provisions affecting the federal features, ratification by the legislatures of not fewer than half the States. The ordinance power borrows only the ordinary legislative competence of Parliament, and an instrument promulgated in a recess cannot perform the Article 368 process or any part of it. A change to the constitutional text that a government needs must therefore wait for a session and be carried through Parliament in the amendment form, however urgent the government considers it.

Q: Can an ordinance impose or change a tax?

Yes. Because the power is coextensive with Parliament’s legislative competence, an ordinance may impose, vary, or abolish a tax, and may amend taxing statutes, as an Act may. The Supreme Court upheld that width in R. K. Garg v Union of India in 1981, rejecting a challenge to a fiscal scheme carried by ordinance. The limit lies elsewhere: the appropriation of money out of the Consolidated Fund of India must follow the procedure of Articles 114 and 266, which runs through a vote of the Lok Sabha and an Appropriation Act, and the ordinance route does not displace that procedure. Tax law may change in a recess; the commitment of public money may not be completed there.

Q: Can an ordinance operate retrospectively?

Yes, within the limits that bind retrospective legislation generally. An ordinance is legislation, and like an Act it may be given retrospective operation in civil, fiscal, and regulatory matters, and may validate past action by altering the legal basis on which that action was taken. The hard exception is criminal: Article 20(1) forbids conviction under a law that was not in force when the act was done, and forbids a penalty greater than the law then allowed, and that prohibition binds an ordinance exactly as it binds an Act. Retrospective ordinances are a standard response where a tax demand or a regulatory scheme has failed in court on a curable defect of the underlying law.

Q: Can the President withdraw an ordinance, and at what stage?

Yes, at any time. Article 123(2)(b) provides that an ordinance may be withdrawn at any time by the President, and Article 213(2)(b) gives the Governor the same power. Withdrawal is available before Parliament reassembles, during the six-week window, and in principle at any point before the instrument ceases to operate on its own. In practice withdrawal, exercised on ministerial advice, is the government’s way of abandoning a measure that has met decisive resistance, of replacing an instrument whose exigency has passed, or of conceding a point ahead of a disapproval debate. Withdrawal ends the instrument from the moment it takes effect; it does not erase the period during which the ordinance operated as law.

Q: Can a Union ordinance legislate on a subject in the State List?

Only where Parliament itself could. The ordinance power borrows Parliament’s competence and no more, so a Union ordinance may enter the State List only through the same gates that admit an Act of Parliament: principally during a proclaimed emergency, where the Rajya Sabha has resolved that legislation on a State subject is necessary in the national interest, where two or more States have requested the legislation, or to implement an international agreement. Outside those gates, a provision on a State List subject in a Union ordinance is a provision Parliament would not be competent to enact, and Article 123(3) declares it void to that extent. The rest of the instrument stands if the offending provision can be separated.

Q: What did the Forty-fourth Amendment change about challenges to ordinances?

It removed the bar that the Thirty-eighth Amendment had placed in the courts’ way. The Constitution (Thirty-eighth Amendment) Act, 1975 had inserted clause (4) into Article 123, and a corresponding clause into Article 213, declaring the satisfaction of the President or the Governor to be final and conclusive and not questionable in any court on any ground. The Constitution (Forty-fourth Amendment) Act, 1978 omitted those clauses, with effect from 20 June 1979. The omission restored the position the Supreme Court had taken in R. C. Cooper in 1970: the satisfaction is examinable on the ground of mala fides, while the adequacy of the circumstances remains for the government and, politically, for Parliament.

Q: Did the ordinance-making power exist before the Constitution of India?

Yes, in a direct line of descent. The Indian Councils Act of 1861 first armed the Governor-General with a power to issue ordinances in cases of emergency, giving them a life of six months. The Government of India Act of 1935 carried the device in section 42 for the Governor-General and in section 88 for the Governor of a province, with the instrument subject to disapproval and to a period running from the legislature’s reassembly. The Constituent Assembly debated the inheritance and retained it in Articles 123 and 213, over warnings from members including H. V. Kamath that the power could be abused, and on Dr B. R. Ambedkar’s defence that circumstances could arise between sessions which could not wait.

Q: Can an ordinance be struck down for violating fundamental rights?

Yes. An ordinance is expressly included in the definition of law in Article 13(3), and Article 13(2) forbids the State from making any law that takes away or abridges the fundamental rights in Part III, declaring such law void to the extent of the contravention. The Supreme Court confirmed the position in A. K. Roy v Union of India, holding the ordinance to be law for Articles 13 and 21 and subject to the same constitutional limitations as an Act. The temporary character of the instrument is no defence: a right abridged for the months of an ordinance’s life is abridged, and the court may declare the offending provision void from the start.