Every few years the same legislative argument returns to Indian public life in a new costume. A benefit that a government distributes by executive decision, a health cover, a pension, an income support payment, an urban jobs programme, is said to be too important to rest on a guideline, and the demand is made that Parliament enact it. The opposing answer is equally familiar: a statute would freeze what ought to stay flexible, convert a budgetary choice into a legal compulsion, and enrich litigants rather than beneficiaries. Both sides argue about rights-based welfare as though the question were one of principle. It is not. India has run both models at national scale for long enough that the question is empirical: what has statutory backing actually done for the programmes that received it, dimension by dimension, and what has the scheme form actually done for the programmes that stayed inside it? Rights-based welfare is the subject of this article because the comparison, conducted on the record rather than on the theory, produces a verdict narrower and more useful than either advocacy offers.

The comparison that follows is built on the field as the series has mapped it. The complete map of India’s welfare instruments, the constitutional layer, the judicial bridge, the insurance and family-duty models, the rights turn and the scheme layer, belongs to the complete guide to welfare laws in India, which owns the source distinction and the classification of individual programmes; this article takes that classification as given and asks the question the guide routed here, whether the statutory side of the line outperforms the executive side, and on what. The statutory instances examined are the employment guarantee, the food security statute, the education statute, the disability statute and the forest rights statute. The scheme instances are the principal programmes in housing, health cover, income support, nutrition and pensions. The method is the one the brief for any honest comparison imposes: eight dimensions, a verdict recorded for each on the evidence, and the general verdict counted out of the column rather than asserted over it. The claim the counting supports can be stated in advance so the reader can test every section against it. Statutory form buys continuity, not adequacy: on the evidence, giving a welfare programme statutory backing has reliably protected it from withdrawal and has reliably created an enforceable route to a remedy, and has not reliably produced better funding, faster payment or wider coverage, so the case for legislating a programme rests on permanence and on the availability of a remedy rather than on any expectation that the money will follow.

The experiment that makes the comparison possible has a date. For the first half century of the Republic, the statutory side of welfare was occupied by the insurance statutes for the organised workforce and the maintenance duties laid on families, while the subsistence of the unorganised majority was carried, where it was carried at all, by executive programmes that succeeded one another under changing names. The rights turn of the mid-2000s changed the experiment’s design. Within a single decade Parliament converted rural employment, food distribution and elementary education, the three largest subsistence programmes in the country, from administrative programmes into statutory entitlements, recognised forest rights in statutory form, and then, in the following decade, rebuilt disability law as a rights code, while the same governments simultaneously launched the largest health cover and the largest income support payment in the framework’s history as pure executive schemes. Both models were therefore run at full national scale, by the same administrative state, across the same decades of fiscal weather, digitisation and political alternation. The comparison is not between a tradition and a proposal. It is between two bodies of delivered fact large enough to bear a verdict, and the verdict’s narrowness, when it comes, is a property of the evidence rather than of any caution in drawing it.

Rights-based welfare and scheme-based welfare: two models defined by their legal source

A rights-based programme and a scheme-based programme can deliver the identical object, a day of work, a kilogram of grain, a hospital admission, a monthly pension, through the same office, the same bank account and the same official, and remain different legal objects, because everything about the beneficiary’s position is fixed by the instrument that created the programme and by nothing else. The first model rests on an Act of Parliament or of a state legislature. The statute names the beneficiary in its own text, places a duty on a named authority, prescribes the machinery through which the duty is performed and the forum through which a failure can be challenged, and can be altered only by the same solemn route that created it, a bill introduced, debated and passed. The second model rests on an executive decision, given effect through the spending power the Constitution confers in Article 282, approved through the expenditure machinery of the government, and issued to the administration as guidelines by the ministry that owns the programme. No legislature votes on the programme as such. The legislature votes the ministry’s demand for grants in aggregate, and the programme exists inside that aggregate as an administrative commitment, revisable by a revised guideline, an office memorandum or a budget line reduced to silence.

The difference is inscribed in each instrument’s life cycle. A scheme is born in an announcement or a budget proposal, is approved for a period characteristically aligned to the cycle of a Finance Commission, and must be appraised afresh for continuation, at which point restructuring, merger and closure are ordinary administrative events requiring no public proceeding. A statute has no appraisal date. It continues until a legislature repeals it, and repeal is a public act that must be defended on the floor of a House. Between those two life cycles sits every other difference this article measures, and most of them are consequences of this one: the executive that can revise its own instrument at will holds a power over the scheme beneficiary that no executive holds over the statutory beneficiary, however generous the scheme and however rationed the statute.

What does a welfare statute actually change for the person at the counter?

It changes who can say no, and on what grounds. Under a statute, the officer at the counter applies a text the officer’s own government did not write and cannot rewrite: the beneficiary class, the duty and the forum are fixed by the legislature, and a refusal must be justified against that text.

For the person at the counter, the statute’s first gift is not money but a fixed point. The eligibility written in an Act can be argued with, because it sits in a public text that the beneficiary’s adviser can read as easily as the officer can. The eligibility written in a guideline can also be read, but the guideline carries its own amendment power inside it, and the version that governs a claim is the version in force when the claim is decided, which the issuing authority may have changed since the claim arose. The second gift is a respondent. A statutory duty has a named bearer, a Programme Officer, a District Grievance Redressal Officer, a Commissioner, against whom a proceeding can be framed. A scheme’s duty is owed by a programme, which is to say by everyone in a chain and therefore by no one who can be summoned. The third gift is time. Statutes characteristically attach periods to duties, fifteen days to provide work, a period within which a wage must be disbursed, and the lapse of the period is itself a measurable breach. Guidelines attach targets to programmes, and a missed target is a statistic, not a breach. None of these gifts is a payment. All three are preconditions of payment being claimable rather than merely expected, and the sections that follow test how far claimability travelled into delivery.

The first dimensions: source and the nature of the claim

The legal source decides the nature of the claim, and the nature of the claim is the dimension on which the two models differ most completely. A statutory entitlement is a claim of a defined person against a defined duty-bearer, arising on defined facts: a rural household whose adult members volunteer for unskilled manual work, a child between six and fourteen, a person with benchmark disability, a household inside the food statute’s coverage share. The claim exists before any official acts, because the Act’s operative verb creates it, and the official’s role is to recognise a position the text has already conferred. Eligibility under a scheme is a claim to be considered for an allocation. The guideline defines a class that may receive the benefit if the programme’s funds, targets and lists reach them, and the individual’s position ripens into anything resembling a right only at the moment of sanction, which is an administrative act the programme performs rather than a status the text confers.

This is not a distinction of drafting taste. It determines what an error looks like. Where a statutory beneficiary is wrongly excluded, the exclusion is an illegality: the officer has misapplied a text that binds the officer, and the correction restores something that was the claimant’s all along. Where a scheme applicant is wrongly excluded, the exclusion is a grievance: the programme has departed from its own announced criteria, the correction is an act of administrative regularity, and the claimant recovers a place in a queue rather than a thing withheld. Courts police the second kind of error, under the doctrine associated with Ramana Dayaram Shetty v International Airport Authority of India, decided in 1979, that the state distributing its largesse must act reasonably and without arbitrariness, and that doctrine is a real protection while a scheme subsists. It is also a protection with a defined edge. It attaches to the administration of the scheme and not to its existence, it polices departure from the guidelines and cannot prevent the guidelines from being changed, and it therefore secures the beneficiary’s treatment inside the programme while leaving the programme itself, and every future benefit under it, exactly as insecure as the executive chooses to leave them.

The coverage dimension belongs with the claim, because a claim is only as wide as the class the instrument defines. Under the statutory model, coverage is determined by statutory criteria: the definition of the beneficiary sits in the Act, and narrowing it requires amendment. Under the scheme model, coverage is determined by executive criteria, and narrowing it requires a circular. The recorded behaviour of the two models under this dimension is more equivocal than the drafting difference suggests, and the equivocation is the subject of a later section: statutory criteria have proved capable of freezing coverage as effectively as any guideline, because a coverage percentage applied to a population count that is never updated narrows in real terms every year without any amendment at all. The drafting difference is real, the freezer is real, and the verdict column at the end of this article records both.

Two refusals: the claim tested at the counter

The dimensions become concrete when they are followed through a single refusal under each model, and the two refusals that follow are assembled from the machinery the delivered record documents rather than invented for effect. A rural household applies for work under the employment guarantee in the lean season. The application is taken at the panchayat, and in the programme’s documented weakness the dated receipt that starts the statutory clock is not issued, or the work that opens is a fraction of the days demanded, because the approved labour budget for the quarter is spent. The household’s legal position, even in this degraded instance, has a definite shape. The Act entitles the household to demand work, fixes fifteen days as the period for providing it, prices the failure in an unemployment allowance, and names the officers, the Programme Officer and above that office the District Programme Coordinator, to whom the demand and the complaint are addressed, with a district Ombudsman appointed under the programme’s instructions sitting over the structure. The household’s quarrel is with the record: whether the application was made, whether the receipt exists, whether the days offered match the days demanded. Every element of the quarrel is capable of proof from documents the statute’s design generates, and the administration’s failure to generate them is itself a contravention the Act names. The claim may fail in practice, for the reasons the capped-entitlement section gives, but it fails as a claim, against a standard, in a forum the programme does not control.

A pensioner under the assistance programme is omitted from the payment list in the same season, because a revised verification exercise has struck the name, or because the state list from which the central assistance flows has been redrawn against a census base, or because the installment has simply not been released. The pensioner’s position has no corresponding shape. No Act defines the pensioner. No officer owes the pensioner a duty within a period any text prescribes. No allowance accrues for the months of omission, and no forum outside the programme’s own administration exists to which the omission can be carried as a legal wrong. The pensioner’s remedies are an application for correction inside the programme’s procedures, a complaint through its channels, the intervention of a representative, and, at the limit, a writ petition arguing that the administration has departed arbitrarily from its own guidelines, an argument that, if it succeeds, restores the pensioner to the list the guidelines describe and gives no protection against the guidelines being redrawn in the next appraisal. The two refusals are identical in human terms: a poor household, a benefit the state announced, a counter at which the benefit stopped. Everything that differs between them is legal architecture, and the architecture is the whole of the difference between a benefit that can be demanded back and a benefit that can only be asked for again.

The illustration also locates the statutory model’s advantage with a precision the dimension table cannot. The advantage does not lie in the first refusal ending happily; on the programme’s record it often does not. It lies in the identity of the questions the refusal generates. Under the statute, the questions are about the administration’s compliance with a standard it did not set: was the application recorded, was the period honoured, was the allowance computed. Under the scheme, the questions are about the administration’s fidelity to a standard it set itself and may reset: was the name on the list, did the guideline in force cover the case, was the revision prospective. A beneficiary’s adviser reading the two files knows at once which file can be argued upward, to a text, a forum and a court, and which can only be argued inward, to the programme’s better nature and the political cost of its obduracy. Litigation strategy, grievance drafting and even the tone of an application at the counter all follow from that single difference, which is why the comparison treats the nature of the claim as the dimension on which the models are furthest apart, ahead even of the funding dimension on which the argument is usually conducted.

The amendment record: how welfare texts actually change

The modification dimension is usually argued in the abstract, statutes hard to change, guidelines easy, and the record of how the two kinds of text have in fact changed adds a qualification in each direction. Statutes in this field have not been immobile. The employment statute was renamed for Mahatma Gandhi by amendment in 2009, a symbolic change carried through the full legislative route for a programme whose operative provisions were untouched. The education statute was amended in 2019 to restore the possibility of detaining a child at the fifth and eighth classes after a re-examination, reversing one of the original Act’s most debated design choices, the no-detention policy, after a decade of argument about learning outcomes in which state governments were the principal complainants. The disability statute of 1995 was not amended into its successor at all but repealed and replaced in 2016, after a convention ratification made the older text’s welfare-and-charity model untenable. Change under the statutory model is therefore real, sometimes substantial, and occasionally a reversal. What distinguishes it is not its rarity but its form: each change was a bill, debated in public, reported on, voted, and recorded in a text whose before and after versions any citizen can compare. The beneficiary whose position a statutory amendment worsens at least confronts a change that had to survive a legislature, and the political cost of the proceeding is itself a protection, since provisions that cannot survive debate tend not to be moved.

Guideline change has the opposite profile in every particular. It is frequent, unrecorded in any single public instrument, and frequently invisible to the beneficiaries it affects until the effect arrives. A funding ratio between the Union and the states is revised in an office memorandum and a programme’s economics alter in every district. A verification drive revises a beneficiary register and names fall off lists for reasons recorded, if at all, inside the administration’s own files. An appraisal at the end of a Finance Commission cycle restructures a programme’s components and the beneficiary discovers the change at the counter. None of this implies bad faith; much guideline revision is the ordinary maintenance of a large administrative state, and some of it corrects errors a statute would have preserved for years. The comparison’s point is narrower and is the one the verdict column records: the two instruments distribute the burden of change differently. Under a statute, the government that wants change must carry the burden of persuading a legislature in the open, and the beneficiary’s existing position is the default that change must displace. Under a scheme, the beneficiary who wants the position preserved must carry the burden of persuading the administration, and change is the default that preservation must displace. Defaults decide most contests in which one side does not fight, and welfare beneficiaries, dispersed and poor, are the class least able to fight every contest. The modification dimension scores for the statutory model not because statutes never change but because the statutory default is set in the beneficiary’s favour, and setting the default is, over decades of administrative weather, most of what legal form does.

Justiciability and grievance machinery: the remedy dimension

A statutory entitlement is enforceable by a court in the direct sense: the beneficiary pleads the Act, proves the facts that bring the case inside its definitions, proves the denial, and asks for the duty to be performed, and the court that grants relief is applying a text rather than designing a policy. A scheme benefit is reviewable only on administrative law grounds: the court may strike down an arbitrary exclusion, compel the administration to follow its own guidelines, or condemn a classification that offends the equality guarantee, and it will ordinarily stop there, because the content of the programme, its rates, its coverage and its continuation, are matters of policy into which the judicial role does not extend. The grievance dimension runs parallel. The rights-turn statutes characteristically build a forum into the text, a grievance redressal officer in the establishment under Section 23 of the disability statute, a District Grievance Redressal Officer and a State Food Commission under the food statute, Maintenance Tribunals under the senior citizens statute, together with Chief and State Commissioners for disability with powers of inquiry, so that the beneficiary’s first remedy sits inside the statute and near the counter. The enforceable-remedy sentence of this comparison is the one the disability cluster demonstrates in practice: under the reservation and accessibility provisions of the disability statute, a person with benchmark disability denied a reserved seat or an accessible building holds a text, a named officer and a Commissioner to go to, a position no beneficiary of any health, housing or pension scheme in the country occupies. Schemes offer departmental channels instead: a helpline, a complaint portal, an appeal to the officer above the officer who decided, each useful for correcting an error the programme concedes and none capable of entertaining the complaint that matters most, that the programme itself has been narrowed, delayed or closed.

Why do courts treat a statutory entitlement differently from a scheme?

Because the two cases ask the court to do different work. A statutory case asks the court to apply a text the legislature has already written, so the order performs the legislature’s decision. A scheme case asks the court to police the manner of a policy decision, its fairness and its fidelity to the guidelines, not its content.

The practical weight of the grievance dimension needs one qualification, which the field guide records and this comparison adopts: the statutory forums are real, formally complete and lightly used. Beneficiaries frequently do not know the forum exists, several of the bodies depend on state action for their constitution and exist unevenly across the country, and the officers who staff them are drawn from the administration whose decisions they review. A remedy that is never invoked delivers no more than a helpline does. The dimension still scores for the statutory model, for a reason the usage figures do not capture: the existence of the forum changes the behaviour of the counter even when the forum is never approached, because an officer who knows a refusal can be carried to a named authority with power to direct correction refuses differently from an officer whose refusal can be carried only upward through the officer’s own chain. Deterrence is a delivery mechanism, and it is one the scheme model cannot replicate, since a departmental channel reviews the officer by the officer’s own standards.

The funding dimension: a statutory charge in some instances, an appropriation in all

The funding obligation is the dimension on which the popular argument for statutes is weakest and the record is clearest. No welfare entitlement in India is self-executing at the treasury. Every rupee under both models is voted annually, and the appropriation is the point at which the two models meet: the employment guarantee’s wage bill, the food subsidy’s vast standing charge behind the food statute’s issue prices, the pension scheme’s budget line, all pass through the same annual vote of the legislature on demands for grants. What a statute can do, and a guideline cannot, is fix the division of the burden and the conditions of the flow. The employment statute writes its funding pattern into Section 22, the Union bearing the entire unskilled wage cost and three-fourths of the material cost, the state bearing the remainder and the unemployment allowance, and that division cannot be revised by a budget circular. The food statute obliges the Union to allocate the foodgrains the entitlement requires and structures the sharing of the distribution system’s costs. A scheme’s funding pattern, by contrast, a ratio between the Union and the states fixed by executive decision for centrally sponsored programmes, is policy from beginning to end, revisable by the same route that set it, and a state that delays its share stalls the programme inside its territory without breaching any text.

What neither instrument can do is fix the total. The statutory funding share binds the proportions of whatever sum is appropriated; it does not appropriate the sum. A legislature that enacts an uncapped entitlement and then votes a fraction of its full-demand cost has not broken the statute in any way a court can easily reach, because the duty to provide and the duty to fund travel through different constitutional channels, and the second is performed, if that is the word, in the estimates of a single year. The employment guarantee’s record is the standing exhibit. People’s Action for Employment Guarantee, the network of organisations and researchers that tracks the programme, has estimated that roughly a fifth of the funds available in a year are absorbed by clearing the previous year’s unpaid wage and material liabilities, so that the nominal allocation overstates the year’s working resources before a single muster is opened, and allocations have repeatedly been exhausted well before the year’s end. The scheme model suffers the identical constraint without the statutory share: a scheme that is not funded in a given year is not breached but merely discontinued in practice, and no arrears accrue against it in any ledger a beneficiary can point to. On funding, therefore, the two models differ in the architecture of the obligation and not in its reliability, and the verdict column records no discernible difference, because the evidence shows statutory programmes rationed by appropriation exactly as schemes are rationed by it, with the single asymmetry that the statutory shortfall is visible and attributable while the scheme shortfall is absorbed invisibly into a smaller programme.

The education statute displays the funding dimension inside a single provision. Its central admissions obligation requires private unaided schools to admit children from weaker sections and disadvantaged groups to at least twenty-five per cent of the entry-level class, and the state reimburses the school for each such child at a rate the statute ties to the state’s own per-child expenditure or the fee the school charges, whichever is less. The entitlement is complete in the text; the money that makes it acceptable to the institutions that must honour it travels by reimbursement, in arrears, through state budgets of varying capacity, and the working of the twenty-five per cent quota in practice is substantially a record of that flow, of seats filled or left empty in the rhythm of a payment that arrives after the service has been rendered. A legislator reading the episode learns the funding dimension’s whole lesson in miniature: the statute settled who must be admitted, the appropriation decided how gracefully the settlement would operate, and no amendment to the admission duty could have repaired a reimbursement delayed by a treasury.

One refinement of the funding dimension belongs here because it explains how a statutory price can be maintained for years while the entitlement behind it is rationed. The food statute fixes its issue prices in a Schedule, three rupees for a kilogram of rice, two for wheat and one for coarse grains, and the difference between those prices and the economic cost of procuring, transporting and storing the grain is borne by the Union as food subsidy, an annual appropriation standing behind the statutory text. The arrangement shows the two models’ funding interlocked inside a single transaction: the beneficiary’s side of the counter is governed by the statute, which fixes the price and the quantity as entitlements, while the state’s side is governed by the budget, which must find the subsidy the transaction generates, and any failure of the budget expresses itself not at the price, which the text holds, but in the system’s other joints, in procurement, in stocks, in the grain’s quality and movement and in the arrears owed along the chain. A scheme inverts the expression: its budget failure appears at the counter itself, as a reduced rate, a delayed installment or a shortened list, because no text holds any element of the transaction steady. The beneficiary of the statute is therefore protected in the terms of the benefit and exposed in its logistics, and the beneficiary of the scheme is exposed in both, which is a real difference of degree inside a dimension the verdict column still scores as tied, because the total each model delivers remains, in both, whatever the year voted.

The capped entitlement: where the distinction weakens in operation

The strongest evidence against the simple form of the statutory case is supplied by the statutory programmes themselves, and an honest comparison has to state it at full strength. Several statutory entitlements are in practice capped by allocation, so that the distance between an uncapped entitlement and an allocated benefit, which is absolute in law, is modest in operation. The food security statute’s coverage has been anchored to a population figure that has not been updated: Section 9 of the Act requires the number of persons covered in each state to be calculated on the population estimates of the census whose figures have been published, the census in force at the Act’s commencement was that of 2011, and the percentages the Act fixes, up to seventy-five per cent of the rural population and fifty per cent of the urban, therefore yield a ceiling of some 81.35 crore persons computed on a population the country has since outgrown. No amendment froze the entitlement. No guideline narrowed it. The text itself, applied to a census that a delayed enumeration has left unreplaced, holds coverage at a count that each passing year makes smaller in proportion to the population it was meant to serve, and a person born into an uncovered household in the interval holds a statute that cannot see them. The employment guarantee’s demand is rationed at the point of application by the mirror mechanism: the programme plans through an annual labour budget in which districts project demand and the Union approves the projection as the basis of funding, applications are recorded unevenly and sometimes not recorded at all, and work is opened within the envelope the approved projection and the releases create, so that the household’s uncapped right to demand one hundred days meets, at the counter, a supply of days that was fixed in advance by an estimate. The two records are documented from inside their own statutes by the food statute’s distribution record and the employment guarantee’s implementation record, and between them they establish the capped-entitlement finding on which this comparison’s verdict depends: the distinction between an entitlement and an allocation is weaker in operation than in law, because the appropriation and the register together can do to a statute, quietly and without illegality, much of what a revised guideline does to a scheme.

How can an entitlement be rationed without any amendment to the Act?

By acting on the three things the Act does not fix: the size of the annual appropriation, the projection through which funds are released, and the register through which the defined beneficiary becomes an identified one. The text stands untouched and the duty unamended, while the delivered benefit contracts inside both.

The allowances tell the same story from the remedy side. The employment statute prices its own non-delivery: an unemployment allowance where work is not provided within fifteen days of application, and compensation at a prescribed rate where wages are delayed beyond the statutory period. Both sums are entitlements in the fullest sense, computed by formula from dates the system records, and both are paid, on the programme’s own record, in amounts that are trivial against the liability the formulae describe. The implementation record sets out the mechanism of the failure: the allowance accrues only on a dated and recorded application, and the recording is performed by the administration that would owe the allowance; delay compensation is computed by the programme’s information system on only the first of the payment chain’s two stages, so the share of delay attributable to the second stage is never converted into a figure at all. A statutory remedy can therefore be defeated without defiance, by the design of the instrument that measures it. The comparison must carry this finding into the verdict honestly: the statutory model creates the enforceable route, and the route’s surface is maintained by the same administration whose failures the route exists to correct, so the route is real and its gradient is uphill.

Coverage counted two ways: the register decides

The coverage dimension deserves its own evidence, because it is the dimension on which the statutory model’s legal superiority and its practical performance diverge most widely, and the divergence has a single mechanism: identification. A statute determines coverage by criteria in its text, but no text identifies a single beneficiary. Between the defined class and the paid benefit stands a register, and the register is built by an administrative exercise under both models alike. The food statute fixes the shares and leaves the names to the states, working from the Socio-Economic and Caste Census and from inclusion and exclusion criteria the states frame. The housing and health-cover schemes draw on the same census base and its deprivation criteria. The disability statute gates its reserved benefits behind a medical authority’s certificate and a threshold of forty per cent. The employment guarantee substitutes self-selection for a register of the poor, and then builds a register of job cards anyway, from which names can be and are deleted in verification drives. In every case the operative coverage of the programme, the set of human beings who can actually present the claim, is the product of the register, and the register is an executive artefact whatever the source of the programme above it.

The consequences for the comparison run in both directions and the verdict column’s tie on this dimension is earned rather than diplomatic. In the statutory model’s favour, the criteria the register must implement are fixed outside the administration’s discretion, so an identification exercise that departs from the text is correctable against it, and the grievance machinery of the rights statutes is occupied to a great extent with exactly such corrections: the household wrongly excluded, the certificate wrongly refused, the classification wrongly made. The definition in the Act gives the excluded person a standard to appeal to that the excluding officer did not write. In the scheme model’s favour, a register built for a scheme can be rebuilt at the speed of a new census or a new survey, and several scheme registers have been refreshed through exercises that a statutory coverage base, anchored to a census the statute names, cannot match without the census itself being taken. The food statute’s frozen ceiling is the extreme case of a general phenomenon: statutory coverage criteria are stable because they are hard to change, and they are hard to change in both directions, against the administrator who would narrow them and against the demography that has outgrown them. Executive criteria are unstable in both directions for the mirror reason. A comparison that scored coverage for the statute on the strength of its text, or for the scheme on the strength of its refresh rate, would be scoring the intention of each instrument rather than its record. On the record, exclusion errors, frozen bases and defective registers appear under both, the remedies for them are stronger under one, the refresh is faster under the other, and the population actually reached is decided in both cases by the quality of an exercise neither instrument’s form guarantees.

Delivery performance: the dimension neither model wins

If statutory backing improved delivery, the payment record of the framework’s best-measured statute would show it, and the employment guarantee’s record shows something more interesting: delivery performance tracking administrative design and funding, with legal form nowhere visible as a variable. The implementation record of the guarantee, documented in this series from the programme’s own transaction data, divides a wage’s journey into two stages, the state’s eight days from muster closure to a signed fund transfer order and the Union’s further seven days to credit in the worker’s account. The programme’s own information system computes delay on the first stage alone, and on that measure the Ministry of Rural Development has reported the overwhelming majority of payments as generated in time, a figure that rose through the programme’s second decade under measurement, standard operating procedures and judicial pressure. Independent researchers working from the same transaction records, Rajendran Narayanan, Sakina Dhorajiwala and Rajesh Golani of LibTech India among them, computed the whole journey for the financial year 2016-17 and found twenty-one per cent of payments completed within the statutory fifteen days, with the second stage, the Union’s own, consuming the greater time. Both figures are accurate on their bases. The lesson for this comparison is not the familiar one about the measurement boundary, though the boundary matters to the compensation question. It is that a statutory programme’s punctuality was achieved, where it was achieved, by an electronic chain, a staging discipline and a court’s supervision, instruments available in identical form to any scheme, and was defeated, where it was defeated, by a funding queue that no feature of the Act’s design could dissolve.

A second delivery observation cuts the same way and is worth recording because it disciplines both advocacies at once. Every programme in the framework, statutory or scheme, delivers a visible core that its systems measure and a penumbra of entitlements or service standards that no system counts. The employment statute’s worksite conditions, drinking water, rest and shade, first aid, a crèche where young children are present at the site, exist in a state of statistical invisibility: honoured or not at the individual worksite, recorded in no national series, and therefore absent from every dashboard on which the programme’s performance is defended or attacked. The anganwadi’s service standards under the nutrition programme occupy the identical position on the scheme side, prescribed in the programme’s norms, experienced daily by the children the centre serves, and unmeasured by any instrument that reports the programme’s reach in centres opened and rations distributed. In both models the measured core improves under measurement and the unmeasured penumbra does not, and the beneficiary’s lived quality of the programme is substantially the penumbra. Legal form does not select for measurement; measurement is an administrative decision available to, and neglected by, both models in the same way, and a reformer choosing between the models on delivery grounds will find the choice empty until the measurement question, what the system counts and where its clock stops, has been answered for the programme in hand.

The scheme side of the delivery comparison is correspondingly better than the statutory advocacy predicts. The income support programme for farmer families, a pure executive scheme with no statutory text anywhere in its ancestry, delivers its installments by direct transfer into bank accounts on a published rhythm, and its administrative performance, measured by the regularity with which a defined list is paid a defined sum, compares favourably with the wage chain of the framework’s flagship statute. The health cover scheme processes hospital claims through a national authority’s transaction system with its own timelines and audit trails. Nothing in either programme’s delivery depends on a right, and nothing in the employment statute’s delivery was secured by one: the wage arrived late despite the compensation clause, and the pension installment arrives on schedule despite the absence of any clause at all. The variables that explain delivery, on this evidence, are the simplicity of the benefit, the quality of the register, the directness of the payment channel and the adequacy of the appropriation, and legal form is not among them. That is the evidentiary basis of the table’s tied rows on funding and coverage and of the verdict’s central refusal: a legislator who enacts a programme in order to make it pay better is buying a product the statute does not contain, and an administrator who defends a scheme by its payment record is defending a virtue the scheme form did not produce. Delivery is an administrative achievement. Continuity and remedy are legal ones. The comparison holds because the two kinds of achievement have been kept separate, and most public argument on the subject consists of trading one for the other.

The continuity record: what has never been withdrawn

Against the capped-entitlement finding stands the countervailing finding, and it is equally a matter of record rather than of theory. No statutory welfare entitlement in India has been withdrawn. The employment guarantee has survived changes of government at the Union, public criticism from within administrations that inherited it, and repeated predictions of its dilution, and it operates, amended at the margins, on the architecture its Act laid down in 2005. The food statute converted a distribution system that governments had redesigned at will for decades into an entitlement whose coverage percentages and issue prices sit in a text, and the system has been reformed, digitised and made portable inside that text rather than by replacing it. The education statute’s admissions obligation survived constitutional challenge and continues. The forest rights statute’s recognition process continues. Even where a statutory entitlement has been replaced, replacement has taken the legislative form: the disability statute of 1995 was repealed and replaced by the Act of 2016 through a full parliamentary proceeding that expanded the recognised categories and the reserved shares, so that the beneficiaries of the old text were carried into a new one by a public act whose terms were debated, not extinguished by a circular whose terms were announced.

The scheme record runs the other way, and its pattern is so regular that it can be stated as the layer’s life cycle. The rural employment programmes that preceded the 2005 Act succeeded one another by replacement: the National Rural Employment Programme and the Rural Landless Employment Guarantee Programme were merged into the Jawahar Rozgar Yojana, which was itself succeeded and folded into the Sampoorna Grameen Rozgar Yojana, which the statutory programme then absorbed, and across that entire sequence no worker acquired, at any point, a claim to a single day of work that survived the next restructuring. The rural housing assistance that ran for decades as the Indira Awaas Yojana was restructured and renamed as the Pradhan Mantri Awas Yojana, with revised criteria and a revised beneficiary list drawn from a newer census base, so that households sanctioned under one instrument’s list waited for a successor instrument to recognise them. The health insurance scheme for poor households that operated as the Rashtriya Swasthya Bima Yojana was subsumed into the Ayushman Bharat health cover, its enrolled population carried into a programme with different criteria, a different administering authority and a different benefit design. Merger and renaming are not abuses of the scheme form; they are its normal metabolism, the means by which a new administration marks a programme as its own and by which criteria are updated without legislative delay. Their cost is paid entirely on the beneficiary’s side of the counter, in the currency the statutory model exists to supply: the assurance that the instrument under which a family plans its year will still exist, on the same terms, when the year arrives.

Can a beneficiary sue when a welfare scheme is discontinued?

No, not against the discontinuation itself. A court will review the manner of a scheme’s administration while it subsists, and can strike down an arbitrary exclusion or compel the administration to follow its own guidelines, but the decision to narrow, merge or close a programme is a policy decision no court will convert into a beneficiary’s claim.

One qualification keeps the continuity finding honest, and it cuts in the scheme model’s favour. Continuity by law is not the only continuity there is. The National Social Assistance Programme has paid its pensions since 1995, across governments of every composition, without any statutory protection at all, sustained by the political weight of a large and sympathetic beneficiary class and by the electoral cost of touching it. A scheme with millions of organised or visible beneficiaries acquires a security of fact that approximates the security of law, and a comparison that ignored the point would overstate what enactment adds for the largest programmes. The approximation is still an approximation. The assistance programme’s rates, age bands and categories have been revised by guideline within that continuity, its central contribution to the old age pension has stood for long periods at levels set by administrative decision rather than by any index, and the pensioner’s security has been security of the programme’s existence, not of its content. The statute protects both. The politically protected scheme protects the first and leaves the second to the same discretion that created it, which is the continuity dimension’s precise verdict: the statutory model performs better, and its margin is the margin between a text and a mood.

The courts: the asymmetry that decides cases

The judicial record is the strongest single body of evidence for the statutory model, and the brief for this comparison directs that it be presented as such. Courts have entertained and granted relief in respect of statutory entitlements across the whole range of what relief can do. They have directed the implementation of a statute’s own machinery, as the Supreme Court did in Swaraj Abhiyan v Union of India, decided in 2016, where the drought proceedings were measured against the food statute’s provisions and the employment statute’s obligations, and the governments were directed to constitute the State Food Commissions and appoint the grievance officers the food Act requires and to honour the employment Act’s disciplines in the affected districts. They have enforced the payment obligations of the employment statute against both levels of government together, holding in the same proceedings that the burden of compliance cannot be passed between the Union and the states. They have awarded compensation for the failure of a statutory public service, supervised delivery through continuing mandamus over years, and treated the statutory allowance and the statutory timeline as the measure of the state’s duty rather than as aspirations the state had set itself. In each case the court’s work was possible because a text existed that fixed the standard in advance: the judge compared the record against the Act, and the Act, not the judge, supplied the programme.

Against a scheme, the same courts have generally declined to direct the creation or continuation of the programme, treating scheme design as a matter of policy. The deference has a considered form. In State of Punjab v Ram Lubhaya Bagga, decided in 1998, government employees challenged a revised policy that fixed the rates at which medical treatment would be reimbursed, and the Supreme Court upheld the revision, reasoning that no state can have unlimited resources to spend on any of its projects and that the provision of facilities has to be to the extent finances permit, so that the fixation of a rate and scale under the new policy could not be held to violate the right to life or the directive principle on public health. The case is the scheme model’s constitutional portrait: the court protects the beneficiary against arbitrariness, examines the policy for reasonableness, and then returns the content of the benefit, its scale, its rates and by extension its existence, to the authority that made it. No corresponding deference is owed to the administration of a statute, because the content there is the legislature’s, and an officer who revises it downward by practice is not exercising a policy discretion the court must respect but departing from a text the court must enforce.

One episode complicates the asymmetry and must be recorded, because it is the scheme model’s best judicial hour. In the proceedings begun by the People’s Union for Civil Liberties in 2001, against the background of drought and accumulated food stocks, the Supreme Court treated the beneficiaries of the government’s own food and nutrition programmes as holders of entitlements under the right to life, directed that cooked midday meals be served in government and government-assisted primary schools rather than dry rations, appointed commissioners to supervise compliance, and kept the proceeding open for years. Executive programmes were, in that litigation, hardened into judicially supervised obligations without any statute beneath them, and the food statute that arrived in 2013 legislated for a field whose expectations the litigation had formed. The episode shows what the judicial bridge can do at its strongest. It also shows the bridge’s limits as a substitute for enactment. The supervision rested on the right to life rather than on the programmes’ own terms, it depended on a proceeding that a litigant had to sustain and a bench had to keep open, and its directions ran to the operation of programmes the government had chosen to run. It could police a scheme into something resembling an entitlement for the classes and the years the proceeding covered. It could not, and did not, write the coverage percentages, the issue prices or the grievance machinery that the statute later supplied, and a beneficiary outside the proceeding’s classes held nothing from it. Judicial hardening is a bridge built case by case. Enactment is the road.

The asymmetry’s deepest consequence is the distribution of the argumentative burden, which decides more cases than any doctrine. The statutory beneficiary who reaches a court does not have to persuade it that the benefit ought to exist, only that it exists and was denied; the text carries the first and larger half of the argument before the hearing opens. The scheme beneficiary must carry both halves, and the first half is the one courts will not decide. That is why the enforceability difference, which reads as a technicality in a comparison table, is in practice the difference between litigation that enforces programmes and litigation that comments on them, and why the verdict column scores justiciability for the statutory model without hesitation.

Portability and data: the dimension that cuts across both

Portability is the dimension the brief directs to be treated as cutting across both models, and the record justifies the direction. A benefit’s portability, whether it travels with a beneficiary who migrates for work, and the quality of the data systems through which the beneficiary is identified and paid, have been transformed across both layers alike by administrative and technological decisions that owed nothing to the legal source of any individual programme. The ration card’s national portability, under which a food statute beneficiary may draw the entitlement in a state other than the state of registration, was achieved as an administrative and technological arrangement within the statute’s framework rather than by amendment of it. The direct transfer of cash benefits into bank accounts, generalised as administrative policy, removed leakage points and ghost entries from scheme and statute programmes indifferently. The seeding of beneficiary databases against identity records made a scheme payment traceable to a named person with the same completeness as a statutory wage. A beneficiary comparing the two models on the experience of being paid will find the pipe identical, and the pipe is not a trivial part of the experience.

The migrant worker is the dimension’s hardest case and the one in which the two models fail identically for a shared reason. Most instruments in the framework are administered by the state of residence and were designed for a settled population, so the worker who crosses a state line for a season’s work risks falling out of the beneficiary registers of both states: no longer present to be counted at home, not yet resident enough to be registered at the destination. The food system’s answer, national portability of the ration card, was built by administrative arrangement within the statute, and its significance for the comparison is that the statute’s text neither required it nor obstructed it; the entitlement was portable when the administration decided to make it so, exactly as a scheme’s benefit is portable on the same decision. Registration-based instruments display the same neutrality from the other side: a national database of unorganised workers, created by executive action, can carry a worker’s identity across state lines for scheme purposes as readily as any statutory register, and the worker’s difficulty has never been the absence of a database but the absence, at the destination, of a claim the destination’s administration must honour. Portability of the record has been solved as a technical matter under both models. Portability of the obligation, the duty travelling with the person, exists only where a text imposes the duty on whichever administration the person stands before, and in the framework that condition is approximated by the food statute alone, through an arrangement its administrators chose to build. The dimension’s tie is therefore a tie of a particular kind: not two models performing equally well, but one obstacle, the territorial organisation of welfare administration, standing in front of both and yielding to administrative will in either case.

What portability and data cannot do is alter the position when the payment stops. The legal source of a benefit is unaffected by the channel through which it is delivered, and the beneficiary whose transfer fails holds exactly the remedy the source conferred before the channel was built: a claim where the source is a statute, a complaint where it is a scheme. The dimension therefore scores as no discernible difference, with a rider the verdict section carries forward: administrative modernisation has narrowed every gap between the models except the legal one, and by narrowing the others it has made the legal gap the whole of what remains between them. A generation ago the statutory programmes were also, often, the better administered ones, and the comparison confounded legal form with administrative quality. The confound has been removed by the digitisation of the scheme layer, which is why the comparison can be run cleanly, and why its result, that the remaining difference is continuity and remedy, is a finding about law rather than about competence.

The states as the second laboratory

The comparison has been conducted so far at the level of the Union, where the largest instruments sit, but the federation runs the experiment a second time in every state, and the state record sharpens the verdict because the same governments can be watched choosing each model in turn. Maharashtra enacted a statutory employment guarantee in 1977, a generation before the national Act, and the state statute’s two decades of operation supplied the national debate with its working proof that a legal guarantee of work could be administered at scale; the proof existed because a state legislature had been willing to legislate where the Union was still running programmes. Kerala chose the other instrument for the urban gap and has run its urban employment guarantee as a scheme since 2010, gaining the speed and adaptability the model promises and accepting, for its urban workers, every insecurity the model carries. Rajasthan in 2023 carried the statutory form into the field’s newest territory, enacting a statute for platform based gig workers that creates a welfare board and a cess on aggregators, with Karnataka legislating on the same subject after it, so that the constituency whose work the older statutes could not describe acquired its first legal address from a state legislature rather than from a Union scheme. State pension and assistance schemes, meanwhile, pay above the Union’s guideline rates in several states and reach beyond its categories, demonstrating that the scheme model’s generosity, where a state’s politics sustains it, can exceed anything the Union’s statutes offer the same contingency.

The federal evidence matters to the verdict for two reasons. The first is that it separates legal form from the identity of the government wielding it. State legislatures are not more virtuous than Union ministries; the same political class staffs both, and its choices between an Act and a scheme, made within one state’s politics across a few decades, track the selection rule this article derives rather than any party’s doctrine: statute where a constituency organised around a contingency and demanded permanence, scheme where speed or experiment governed. The second is that the state level is where enactment is cheapest to attempt. A state statute binds one state’s administration, can be drafted against one state’s delivery capacity, and generates the demonstration on which a later national enactment can rely, as the Maharashtra guarantee demonstrably did for the Act of 2005. A Union that wishes to test whether a contingency is ready for statutory form therefore has a laboratory the Constitution built for the purpose, and the recurring failure to use it, the leap from national scheme to demanded national statute with no state enactment in between, is a habit of the debate rather than a necessity of the subject. For the beneficiary, the federal structure also means that the audit this series teaches must be run layer by layer: a Union statute, a Union scheme, a state top-up and a state statute may cover the same contingency for the same person, each with its own source and its own remedy, and the person’s real position is the composite, strongest where any layer is statutory and weakest where all four are executive.

The price of enactment, honestly stated

The statutory case is incomplete until its price is stated by someone willing to pay it, and the price has three components that the fiscal debate usually merges. The first is rigidity of the respectable kind: an entitlement written for one decade’s demography and prices governs the next decade until amended, and the food statute’s census anchor is the standing proof that a legislature’s failure to revisit a text is itself a policy, made by omission. Indexation, periodic review clauses and coverage bases tied to the last published census are drafting answers to this component, and their absence from most welfare statutes is a drafting failure rather than an argument against enactment, but the failure is common enough that a legislator must price the risk of enacting a text that no successor will maintain. The second component is the litigation load. An enforceable entitlement invites enforcement, and the writ courts of the country carry welfare litigation in volume, much of it meritorious, some of it strategic, all of it consuming the administrative attention that delivery also needs. The answer that litigation is the sound of a right being used is true and does not refund the attention. The third component is the one finance ministries press hardest: committed expenditure narrows the budget’s freedom, and a state that converts discretionary programmes into statutory charges reduces its ability to respond to a revenue shock by trimming the welfare line, transferring the shock instead to capital expenditure, to borrowing or to the arrears that ration the entitlement from inside. The argument deserves its strongest form, because it is the scheme model’s one genuinely structural defence: flexibility at the level of the whole budget is a public good, and every enactment spends a little of it.

Priced honestly, the three components do not reverse the verdict, and it is worth saying why each falls short of doing so. The rigidity component is a drafting discipline problem: a legislature that writes indexation and review into the text buys continuity of the entitlement’s existence without freezing its quantities, and the censuses and thresholds that froze in the record froze because the texts named fixed bases, not because texts must. The litigation component is partly the remedy dimension viewed from the respondent’s side, and a state that finds its welfare litigation burdensome is invited to consider that the burden is the measurable form of the grievances the statutory design exists to surface, which the scheme model does not abolish but merely renders invisible and unactionable. The fiscal component is the strongest, and it still proves less than its proponents claim, because the budget’s freedom to trim a scheme is exercised, in the record, overwhelmingly against the beneficiaries least able to resist, and a freedom whose predictable incidence is that distribution is not a neutral property of the budget but a standing decision about whose provision is soft. A legislature may still rationally refuse that trade for a given programme, on the selection rule’s own terms. What it may not honestly do is present the refusal as costless, or the scheme’s flexibility as a benefit that someone other than the beneficiary is not paying for.

Eight dimensions, two models, one verdict column

The table below is the findable artifact of this article and the mechanism by which its verdict is derived. Each of the eight dimensions on which the models were compared is recorded with the position under each model and a verdict drawn only from the evidence the preceding sections assembled: whether the record shows the statutory model performing better, the scheme model performing better, or no discernible difference. A reader preparing the comparison for argument or examination can keep the table open as a document through the ReportMedic office file viewer for side-by-side reading of statutes and guidelines beside the texts it summarises.

Dimension	Statutory model	Scheme model	Verdict on the evidence
Legal source	An Act of Parliament or of a state legislature, alterable only by the same route	An executive decision under the spending power, issued as ministry guidelines	Statutory model performs better: the text is public, stable and outside the administering government’s hands
Nature of the claim	An entitlement of a defined person against a named duty-bearer, arising on defined facts	Eligibility for an allocation, ripening only on administrative sanction	Statutory model performs better: an exclusion is an illegality, not a grievance
Justiciability	Enforceable by a court applying the Act, with directions on machinery, timelines and compensation	Reviewable only on administrative law grounds; design and continuation are policy	Statutory model performs better: courts grant relief on entitlements and decline to direct schemes
Grievance machinery	Statutory officers, commissions and tribunals built into the text, beside the writ remedy	Departmental channels inside the programme itself	Statutory model performs better: a forum outside the programme exists, though it is lightly used
Funding obligation	A funding share and conditions fixed in the Act in some instances; the total still an annual appropriation in all	An annual appropriation, with shares fixed by executive decision	No discernible difference: both models are rationed by the same annual vote
Modification or withdrawal	Requires an amending Act in public; no statutory welfare entitlement has been withdrawn	Achievable by revised guideline or budget silence; renaming, merger and closure are routine	Statutory model performs better: continuity of the programme and of its content is protected
Determination of coverage	By statutory criteria fixed in the text	By executive criteria in the guidelines	No discernible difference: statutory criteria are stabler, but a frozen population base narrows statutory coverage in practice
Portability and data	Delivered through the same registries, identity seeding and transfer channels	Delivered through the same registries, identity seeding and transfer channels	No discernible difference: modernisation equalised the pipe and left the legal position untouched

The count is five dimensions for the statutory model, none for the scheme model, and three on which the evidence shows no discernible difference. The shape of the count is the finding, and it deserves a careful reading, because a careless one would score the comparison five to nothing and stop. The scheme model wins no dimension on which the beneficiary’s position is measured, and that is a genuine result: on source, claim, justiciability, grievance and continuity, the statutory form is simply the stronger instrument for the person the programme exists to serve. But the three dimensions on which the models tie are not minor dimensions. They are funding, coverage and delivery, the dimensions that decide how much benefit actually reaches how many people, and on those the statutory form’s superiority in law has not converted into superiority in fact. A comparison that ended five to nothing would be an argument for legislating everything. A comparison that ends five to nothing with the money dimensions tied is an argument for legislating selectively, and the sections that follow draw exactly that conclusion.

Counting the column: the verdict and the deciding factor

The verdict the count supports is the namable claim stated at the outset, and it can be stated with its evidence attached. Statutory form buys continuity, not adequacy. Giving a welfare programme statutory backing has reliably protected it from withdrawal: no statutory entitlement has been repealed away from its beneficiaries, replacement has proceeded by public enactment, and the programmes of the rights turn have outlasted the governments that enacted them and the fiscal weather that followed. It has reliably created an enforceable route to a remedy: a defined claim, a statutory forum, and a court that applies a text rather than deferring to a policy, with the burden of argument inverted in the beneficiary’s favour. It has not reliably produced better funding, since the appropriation governs both models and has rationed the largest statutory entitlement to roughly half its promised days and the largest food entitlement to a population count frozen at a past census. It has not reliably produced faster payment, since the employment statute’s own compensation for delayed wages is defeated by the measurement instrument that was supposed to compute it. And it has not reliably produced wider coverage, since coverage under a statute is hostage to the register and the census base exactly as coverage under a scheme is hostage to the guideline list.

The count also has to be weighted before it is used, because the eight dimensions are not eight equal goods. The five the statutory model wins are the dimensions of the beneficiary’s legal position, and their value is realised in the worst moments of a programme’s life: the year the budget is cut, the season the list is redrawn, the month the payment stops, the day a new administration decides the programme is its predecessor’s vanity. The three tied dimensions are the dimensions of the programme’s ordinary operation, and their value is realised every day. A reader who weights by frequency will find the comparison close, since delivery happens daily and discontinuation happens rarely. A reader who weights by consequence will find it one-sided, since a single discontinuation destroys every year of delivery that depended on it, and a single unremedied refusal in a subsistence programme destroys the household it lands on. The comparison’s weighting rule follows from the selection principle the honest middle derives: weight by the harm the dimension protects against. On that rule the statutory model’s five dimensions are the load-bearing ones for any programme whose interruption maims, and the tied dimensions dominate only for programmes whose benefits are conveniences rather than subsistence, which is another way of arriving at the same selective verdict by counting rather than by intuition.

The deciding factor is therefore not a property of programmes but a question about purpose: what does the legislator want the law to secure? If the object is permanence, protection against the next administration’s indifference and against the quiet revision that no one has to announce, and if the object is a remedy, a route by which a denied beneficiary can compel the state to answer against a standard the state did not write for the occasion, the statute secures both and the scheme secures neither. If the object is adequacy, a benefit large enough, prompt enough and wide enough to meet the need it names, the statute secures nothing by itself, and the legislator who enacts an entitlement and votes a fraction of its cost has purchased the appearance of the first pair of goods while defaulting on the second in a form the courts and auditors can at least name. Separating permanence from adequacy resolves an argument that has been conducted as though the two travelled together. They do not. They are bought with different instruments, the first with legislative form, the second with money and administrative capacity, and a debate that prices them separately will reach better decisions than a debate that treats enactment as a substitute for appropriation or appropriation as a substitute for rights.

Both advocacies overreach

The position that every major welfare programme should be given statutory backing is the stronger of the two public positions and it still overreaches, in a way its proponents rarely price. The binding constraint on most welfare programmes in India is appropriation rather than legal form. Delivery fails, in the main, because the money voted is a fraction of the entitlement’s full-demand cost, because the delivery tier is staffed to a norm that starves it, and because the register through which beneficiaries are identified is built and maintained by the administration whose errors the beneficiary must then correct. A statute that creates an entitlement the budget cannot meet does not abolish those constraints; it routes their effects into different channels, producing rationing at the application counter, arrears in the payment chain, and litigation that establishes the right at the level of principle while the individual claimant waits for a computation no instrument performs. The employment guarantee’s compensation provision is the cautionary exhibit, an entitlement defeated not by repeal or defiance but by a measurement boundary. There is also a cost on the other side of the ledger that the statutory case understates: enactment is slow, amendment is slower, and a programme whose design is wrong in a respect only experience could reveal carries its error in a text until a legislature can be moved to correct it, while a guideline error is corrected by the next circular. Jean Dreze, the economist associated with the design and defence of the employment guarantee and with the right to food campaign, has argued the case for legal guarantees from the record of what discretionary provision did to the poor across the decades of programme churn, and the argument is at its strongest exactly where he makes it, on continuity and on the dignity of a claim over a favour. It does not follow, and the record does not show, that the same instrument improves the funding line of the programme it protects.

The opposing position, that schemes are the more efficient instrument because they are flexible, overreaches in the opposite direction, by describing the same property from the administrator’s end and calling the description a virtue. Flexibility is indistinguishable from insecurity from the beneficiary’s side. The power to revise criteria as evidence arrives is the same power, in the same hands, as the power to narrow a programme between one budget and the next, to merge it into a successor whose list excludes the household the predecessor recognised, and to close it without any proceeding at which a beneficiary could be heard. When a scheme is narrowed, no beneficiary has any remedy at all: not a diminished remedy, not a slow one, none, because the complaint that the programme itself has changed is precisely the complaint the administrative law of schemes cannot entertain. The efficiency claim also fails on its own terms more often than its proponents allow. The delivery evidence does not favour either model cleanly: the best-measured statutory programme pays late through a chain its own system only half observes, and the best-administered schemes pay with a regularity that compares well with badly identified entitlements, so that flexibility has not in fact purchased superior delivery, only superior freedom to discontinue. A virtue that does not appear in the delivery figures and whose cost appears in the continuity figures is not an efficiency. It is an allocation of risk, from the government that designs the programme to the household that plans around it.

What would count as evidence that statutory backing had failed a programme?

Evidence addressed to the dimensions the statute is supposed to win: a programme whose beneficiaries could not reach its forum, whose allowance and compensation provisions existed only in the text, whose coverage had frozen against an unupdated population base, and whose continuity protected an empty shell of the original benefit.

The honest middle: where statutory backing earns its cost

Between the two overreachings sits the position the evidence supports, and it is a position about selection rather than about sides. The strongest case for statutory backing is where the harm from interruption is severe and irreversible. Food is the paradigm: a household’s entitlement to grain is consumed in the month it is denied, the denial cannot be made good by a successor programme’s generosity in a later year, and the beneficiary class is too dispersed and too poor to renegotiate its provision politically each time a guideline is revised. Employment in distress carries the same structure, since the lean season in which the guarantee matters does not recur for the household that migrated or borrowed through it, and education carries it in a third form, because a school year lost to a closed admissions obligation is not recoverable by the child who lost it. In each case the value of the programme is concentrated at the moment of need, interruption destroys the value rather than deferring it, and the statutory form’s specific gifts, continuity and a remedy, are worth their price in rigidity and litigation. The weakest case for statutory backing is the mirror image: a programme that is genuinely experimental, whose beneficiary definition, benefit level and delivery channel are all expected to be revised as evidence arrives, whose failure modes are not yet known, and whose beneficiaries are not exposed to irreversible harm while the design settles. Enacting such a programme freezes its first draft into the hardest text to amend in the legal system, multiplies the forums in which its teething errors can be litigated, and purchases continuity for a design that has not yet earned continuity. The distinction between the two cases is rarely made when the question is debated, because the debate is usually about a government’s virtue or an opposition’s obstruction rather than about the structure of the harm. Made carefully, the distinction decides most real instances: legislate where interruption maims, experiment where revision teaches, and distrust any argument that reaches for the same instrument in both situations.

Two border cases discipline the selection rule from either side. The first is the hybrid statute, the Act that creates the beneficiary class and the fund but leaves the benefits to schemes framed under it, of which the Unorganised Workers’ Social Security Act of 2008 and the Code on Social Security of 2020 are the field’s standing examples, and which the disability statute itself joins in its social security chapter, where the operative direction is cast within the limits of the government’s economic capacity and development. The hybrid purchases the statute’s permanence for the framework and the scheme’s flexibility for the benefit, and a beneficiary under it holds a statutory identity and an inchoate claim. The selection rule reads the hybrid as a confession that the contingency’s benefit design is not yet settled, and judges it accordingly: respectable as a staging post, indefensible as a resting place, because it collects the political credit of enactment while deferring every decision that makes an entitlement real. The second border case is the scheme that a court has hardened, of which the midday meal under the food litigation is the exemplar. Judicial hardening gives the beneficiary a remedy without a legislature, and it can hold a programme in place across the years a proceeding runs, but it reaches only the classes the proceeding covers, rests on the right to life rather than on the programme’s own terms, and lapses as a discipline when the proceeding closes. It is a bridge, and the honest middle treats it as one: evidence that a programme’s beneficiaries need the statutory form, not a substitute for providing it.

The selection rule has one large pending application that the debate returns to more often than the urban guarantee, and it deserves the rule’s application in the open: old age pensions. The assistance programme’s pensions are the largest welfare commitment in the country that rests on executive decision alone, and the case for enacting them satisfies the rule’s first two tests completely. Interruption of a pension inflicts harm that is severe and, for a beneficiary of eighty with no earner in the household, irreversible in the only sense that matters, since the months of destitution are not repaid when a successor guideline restores the rate. The beneficiary class, elderly, widowed and disabled persons below a poverty line, is the least able in the entire framework to renegotiate its provision politically between one appraisal and the next. What the rule’s third test asks, whether the design is settled, is where the honest difficulty sits: the pension’s inadequacy is not a defect of its legal form but of its amount, and an enactment that carried the programme’s existing central rates into a statute would constitutionalise nothing except a forum, while freezing amounts whose real value the guideline at least revises without ceremony. The comparison’s framework resolves the tension rather than merely naming it. Enactment is worth its price for the pension if, and only if, the text carries what the guideline cannot be trusted to carry: an indexed rate or a statutory formula for revising it, a defined class the states’ lists implement against a standard, and a forum in which an omitted name is a legal wrong. A pension statute without indexation would be the hybrid’s error in its purest form, permanence purchased for the framework and withheld from the benefit. A pension statute with it would be the rights turn’s unfinished business completed, and the fact that the distinction can be stated this precisely, in terms of which provisions must be inside the text, is the practical yield of comparing the models on evidence rather than on allegiance.

The live test: an urban employment guarantee

The comparison’s usefulness is tested where the argument is live, and the live legislative demand in this space is an urban employment guarantee. The rural guarantee’s rural boundary is written into its Act, and the urban worker in distress is served, in most of the country, by missions and schemes for urban livelihoods that create no claim to work at all. Kerala has run the counter-experiment since 2010: the Ayyankali Urban Employment Guarantee Scheme, a state scheme named for the social reformer Ayyankali, which promises urban households one hundred days of wage employment in a financial year on local public works, consciously modelled on the rural statute’s design, with job cards, a demand for work, and works executed through urban local bodies. The Kerala experiment demonstrates the scheme model’s genuine strengths at their best. A state government identified a gap the national framework left open, designed a programme for it within its own competence over a Concurrent List field, launched it without waiting for any legislature’s calendar, and tuned it to urban conditions the rural text does not contemplate. It also demonstrates, in its legal character, everything the scheme form cannot give the urban worker it serves. The guarantee in its title is a guideline’s guarantee. Its beneficiary holds no text against the state, no forum outside the programme, no allowance priced against non-delivery by any formula a court could apply, and no protection against the programme’s revision or quiet starvation beyond the political weight of its enrolled households.

What would statutory form secure for such a programme, and what would it not? On this comparison’s evidence, it would secure the claim: a defined urban household’s right to demand work, with a dated application and a period within which work must be provided. It would secure the price of non-delivery, an unemployment allowance whose existence disciplines the administration even where its payment record must be fought for. It would secure the works conditions, the wage and material discipline and the worksite facilities, in a text a ministry circular could not dilute. It would secure continuity across the state’s political cycles, so that the urban poor could plan around the programme as the rural poor have learned to plan around theirs, and it would secure the forum, a named officer answerable for refusal. It would not secure the money: the appropriation would remain an annual act of the same legislature, and a state that enacted the guarantee and starved it would produce the rural programme’s signature failures in an urban setting, rationed applications, delayed wages and an allowance unborn for want of a dated receipt. It would not secure the staffing, the technical capacity to open works at the speed demand requires, or the wage level, which a statute can notify but cannot make adequate against a rising cost of living without an indexation the legislature must itself write in. The proposal therefore deserves a more precise debate than it usually receives. Enactment is the right instrument for the parts of the urban worker’s insecurity that are legal in nature, and an irrelevance to the parts that are fiscal and administrative, and a legislature that enacts the guarantee while writing the funding share, the indexation and the application machinery into the text with the same care as the entitlement itself would have learned the whole of the rural statute’s record, the triumphant half and the cautionary half alike.

The recommendation by reader

The legislator, and the staffer drafting for one, leave with the selection rule and a discipline for applying it. Ask first what interruption would do to the beneficiary: if the harm is severe and irreversible, the programme belongs in a statute, and the drafting care should be spent in proportion on the definition of the beneficiary, the funding share, the forum and the machinery of identification, since those four elements decide whether the enacted right is usable. If the programme is experimental, keep it a scheme without apology, publish the criteria by which its success will be judged, and let the revision the experiment needs happen at the speed of the evidence. In either case, refuse the hybrid that takes credit for enactment while deferring the benefit to schemes, unless the deferral is announced as a stage with the entitlement to follow. The staffer comparing a draft bill against an existing guideline framework, testing which provisions of the programme the text would fix and which it would leave to the executive, does the comparison best with both instruments open at once, and the verdict column of this article’s table is the checklist for that reading.

The advocate, and the adviser assisting a beneficiary, leave with the burden point. Under a statute, plead the text: the definition, the duty, the period, the forum, in that order, and let the respondent carry the argument about resources, which the court will hear as context and not as a defence to a defined duty. Under a scheme, plead the programme against itself: the guidelines as written, the criteria as published, the departure as arbitrary, and frame the claim inside the administration of the programme rather than against its design, because the second framing is the one the court cannot entertain. The general position described in this article is information about the structure of the two models and is not advice on an individual case, where the forum, the limitation position and the facts of the refusal decide the route actually to be taken.

The researcher leaves with the counting discipline the comparison used. Every claim about either model’s performance should be required to state its dimension, its source and its period: delivery figures measured against a programme’s own instrument, continuity measured across changes of government, and coverage measured against the population base the instrument uses rather than the population that exists. The capped-entitlement finding, that a statutory coverage percentage applied to an unupdated census narrows in real terms each year, is a research result waiting to be replicated across the framework’s other frozen quantities, its unindexed rates, its unrevised thresholds and its census-bound shares, and the programme of work it suggests would do more for the statutory case’s credibility than another decade of assertion. The aspirant leaves with the comparison itself, which is directly examinable: the eight dimensions, the asymmetry of the courts, the two findings that qualify each other, and the verdict with its deciding factor, are the skeleton of any answer the civil services papers can frame on welfare legislation, and the candidate who can state the verdict in its narrow form, continuity and remedy rather than adequacy, holds the distinction the question is designed to test.

Where the general verdict does not hold

A comparison article owes its reader the cases that break its own rule, and this one has four. The first is the politically massive scheme. Where a scheme’s beneficiary class is enormous, organised and electorally weighty, the continuity the statute would add is already supplied by politics, and enactment purchases little beyond the remedy dimension; the assistance programme’s quarter century of unstatutory survival is the evidence, and a legislator deciding whether to enact a pension already paid to crores of households should price the statute for its forum and its indexation, not for a permanence the programme already possesses. The second is the statute whose forum is never built. Where the grievance machinery a statute creates depends on state constitution and the state does not constitute it, the statutory beneficiary’s practical position can fall below that of a scheme beneficiary inside a well-run departmental channel, because the formal remedy’s absence also displaces the informal pressure that a programme’s own hierarchy would have applied; the drought proceedings’ need to direct the very constitution of the food statute’s commissions shows the failure is not hypothetical. The third is the contributory model, which the binary comparison brackets and which deserves its bracket to be noticed: the insurance entitlements of the organised workforce rest on statutes, but their security comes from the fund and the contribution record rather than from the entitlement’s form, and a scheme framed under a statute but paid from a dedicated fund can be more secure in fact than a tax-funded entitlement competing in the annual estimates. The fourth is the emergency. Where the need is sudden and the beneficiary class is defined by the event, a flood, a drought, a pandemic’s return migration, the scheme form’s speed is not a convenience but the substance of the response, and the statutory programmes that performed in such moments performed because their machinery already existed, not because enactment is a way of responding to a crisis that has arrived. None of the four displaces the verdict. Each marks a boundary at which the deciding factor, what the law is being asked to secure, returns a different answer, which is what a deciding factor is for.

What would change the verdict

A verdict derived from a count should state what evidence would change the count, and this one’s exposed rows are the three ties. The funding row would move to the statutory model if a statutory programme’s compensation and allowance machinery were ever operated as designed, computed across the whole of the obligation and credited automatically, because a funding system that prices its own failures is a different instrument from one that absorbs them, and no scheme can replicate the discipline without ceasing to be a scheme in the relevant respect. The same row would move, in the opposite accounting, if a major statutory entitlement were repealed or hollowed by an amendment carried against its beneficiaries, since the continuity finding on which two rows rest is a finding about a record that a single future Parliament could break. The coverage row would move if a statutory coverage base were indexed or refreshed by amendment at the pace demography requires, or if a scheme register demonstrated, across a full appraisal cycle, inclusion accuracy that the statutory registers did not match; either result is obtainable, and neither has been recorded. The delivery evidence would begin to separate the models if any large programme changed its legal form while holding its administration constant, the natural experiment the field has never quite run, since programmes that crossed from scheme to statute, the midday meal and the food distribution system among them, changed their machinery and their politics in the same passage. Until such evidence appears, the honest position is the one the count records: the models differ decisively where law operates and barely at all where money and administration operate, and any writer who reports a cleaner result in either direction has simplified the record to fit the argument, which is the failure this article was built to refuse.

Closing verdict

India’s two welfare models are not two philosophies competing for the state’s soul. They are two instruments with different failure modes, and the record of their operation at scale permits the failure modes to be named with precision. The statutory model fails by rationing: the right stands, the definition holds, the forum exists, and the money, the register and the measurement instrument between them deliver a fraction of the promise in a form whose shortfall is at least visible, attributable and challengeable. The scheme model fails by discontinuity: the delivery can be swift, the design can be apt, the payment can be regular for years, and the whole construction rests on a decision that the same authority may revisit without asking anyone, leaving the beneficiary with a complaint where a citizen should have had a claim. Between a model that promises more than it funds and a model that funds only what it promises for as long as it pleases, the comparison’s evidence supports a discriminating preference rather than a camp. Enact where interruption maims and where a remedy matters, which is to say for the subsistence entitlements around which households organise their survival. Administer where experiment teaches, and say so. And in every case, count a programme’s worth on both ledgers the verdict column separates: the ledger of permanence and remedy, which the statute wins, and the ledger of funding, coverage and delivery, on which neither form has yet earned the right to be trusted without audit.

Frequently Asked Questions

Q: What makes a welfare law rights based?

A welfare law is rights based when an Act of Parliament or of a state legislature creates the benefit as an entitlement: the statute defines the beneficiary in its own text, places a duty on a named authority, prescribes a period or a formula by which the duty is measured, and provides a forum, a grievance officer, a tribunal, a commissioner or a court, through which a denial can be challenged. The employment guarantee, the food security entitlement and the education statute are the standard instances. A programme that distributes the same kind of benefit under ministry guidelines, however generous, is not rights based, because the beneficiary holds an expectation under a policy rather than a claim under a text.

Q: Which Indian welfare programmes rest on an Act and which on executive order?

The principal statutory programmes are the rural employment guarantee under the Act of 2005, food security under the Act of 2013, free and compulsory elementary education under the Act of 2009, forest rights recognition under the Act of 2006, disability rights and reservation under the Act of 2016, and senior citizens maintenance under the Act of 2007, beside the older insurance statutes of 1948 and 1952. The principal executive programmes are the National Social Assistance Programme pensions, the Ayushman Bharat health cover, the PM-KISAN income support, the housing assistance programmes and the Integrated Child Development Services. Several programmes are hybrids, resting on a statute that authorises schemes, and the legal source of any named programme should be verified before its position is argued from.

Q: Can a scheme be withdrawn without Parliament?

Yes. A scheme is created by executive decision and financed through the annual budget, and the same authority that created it may narrow, merge, rename or discontinue it by a revised guideline or by the simple absence of a budget line, without any bill, debate or vote on the programme itself. The legislature’s involvement is confined to voting the ministry’s demand for grants in aggregate. Courts will review the fairness of a scheme’s administration while it subsists and can strike down an arbitrary exclusion, but they will not direct the continuation of a discontinued programme, because its design and existence are matters of policy. Withdrawal is therefore lawful, unceremonious and, for the beneficiary, without remedy.

Q: Is a statutory entitlement actually easier to enforce?

Yes, in the specific sense that decides cases. The statutory claimant pleads a text: the definition that covers the claimant, the duty the Act imposes, the period it prescribes and the forum it creates, and asks the court or the statutory officer to apply them, so the argument begins from the proposition that the benefit exists. The scheme claimant must first establish that any enforceable obligation exists at all, and can ordinarily reach only the programme’s own guidelines, arguing that the administration departed from them arbitrarily. Statutory forums are lightly used and statutory allowances are imperfectly paid, so easier does not mean easy; it means the route exists, the standard is external to the administration, and the burden of the larger argument is carried by the Act rather than by the claimant.

Q: Do statutory welfare programmes get better funding than schemes?

Not reliably, on the record of the largest programmes. A statute can fix the funding pattern, the division of cost between the Union and the states, and can oblige the Union to supply the commodity an entitlement requires, and those provisions bind in a way a scheme’s executive funding ratio does not. But the total in both models is set by the annual appropriation, and the appropriation has rationed the statutory programmes as firmly as any scheme: the employment guarantee has delivered on average roughly half its promised hundred days per household across the years its ministry reported to Parliament, and its allocations have repeatedly been exhausted before the year’s end with liabilities carried forward. Statutory form disciplines the architecture of funding. It does not govern its size.

Q: What happens when an entitlement is capped by budget?

The entitlement survives in law and contracts in delivery. The statute is not amended and no refusal is announced; instead the appropriation funds a fraction of full-demand cost, the administrative projection through which funds are released becomes the working ceiling, works are opened and musters issued inside that envelope, and demand that exceeds it is rationed at the application counter, sometimes by the simple failure to record the application at all. Arrears accumulate into the next year, allowances for non-delivery go uncomputed or unpaid, and the beneficiary experiences the cap as delay and shortage rather than as a decision that can be challenged. The Supreme Court has held that the agreed labour budget under the employment statute is not an informal cap on the statutory obligation, which states the law; the delivery record states the practice.

Q: Which welfare model reaches more people?

Neither model owns reach. The largest single coverage in the framework belongs to a statute, the food security entitlement, whose percentages applied to the census of 2011 yield a ceiling of some 81.35 crore persons, while the largest health cover belongs to a scheme, the Ayushman Bharat programme, and the widest pension population is served by the executive assistance programme rather than by any Act. Reach follows the fiscal and political decision to fund a programme at scale, and both instruments can carry such a decision. What differs is the security of the reached population’s position: the person reached by the statute holds a claim at that scale, and the person reached by the scheme holds a place on a list that the issuing authority may revise.

Q: Should more welfare programmes be given statutory backing?

Selectively, on the test this comparison derives. Statutory backing earns its cost where interruption of the benefit inflicts severe and irreversible harm, as with food, employment in distress and elementary education, and where an enforceable remedy matters because the beneficiary class cannot renegotiate its provision politically. It earns least where a programme is genuinely experimental and expected to be revised as evidence arrives, since enactment freezes a first design into the hardest text to amend. The verdict of the delivery record is that a statute reliably buys continuity and a remedy and does not reliably buy funding, coverage or speed, so the case for each new enactment should be argued on permanence and enforceability, and never on the expectation that the money will follow the right.

Q: What is the deciding factor when a government chooses between enacting a welfare statute and launching a scheme?

The deciding factor is what the law is being asked to secure. If the object is permanence, protection against quiet revision and withdrawal by a successor administration, and a remedy by which a denied beneficiary can compel an answer against a standard the executive did not write, the statute is the instrument that secures those goods. If the object is speed of launch, fiscal adaptability or the freedom to redesign as evidence arrives, the scheme is the instrument, and its beneficiary’s insecurity is the admitted price. Adequacy, the size, promptness and width of the benefit, is secured by neither instrument and must be bought separately with appropriations and administrative capacity, which is why arguments that treat enactment as a substitute for funding mislead in both directions.

Q: Can a court order the government to create a welfare scheme?

No, in the ordinary course. The design of welfare provision, what benefit to offer, to whom, at what rate and from what budget, is a matter of policy for the executive and the legislature, and courts have declined to direct the creation of a programme or the continuation of one that has been discontinued, reasoning that provision must be measured against the resources available, as the Supreme Court put it in upholding a revised medical reimbursement policy in State of Punjab v Ram Lubhaya Bagga in 1998. What courts do instead is enforce what exists: a statute according to its text, and a subsisting scheme according to its own guidelines, policed against arbitrariness under the equality guarantee.

Q: What can a court actually order when a statutory entitlement is denied?

The court can order the duty to be performed, and it can reach the machinery as well as the individual case. In the drought proceedings in Swaraj Abhiyan v Union of India, decided in 2016, the Supreme Court directed the implementation of the food security statute’s own machinery, including the constitution of State Food Commissions and the appointment of District Grievance Redressal Officers, and treated the employment statute’s obligations as the measure of the governments’ duties, holding that the burden of compliance cannot be passed between the Union and the states. Orders of this kind are possible because the Act supplies the standard in advance: the court compares the record against the text, directs the officer to be appointed, the allowance to be paid and the timeline to be honoured, and supervises compliance where the failure is systemic.

Q: Why do governments often prefer schemes even though statutes give beneficiaries more security?

Because the scheme’s advantages are real from the government’s side of the counter. A scheme can be announced in a budget speech and be paying beneficiaries within the year, while a bill must survive drafting, cabinet approval, introduction, committee scrutiny in many cases, two Houses and assent, and may emerge amended into a shape its sponsors did not choose. A scheme’s criteria can be tuned to a state’s conditions and its scale matched to the year’s revenues without the embarrassment of amending a statute downward, and a scheme can respond to a newly measured need, a drought, a price shock or a displacement, at the speed of the event. Ministers defending the method make each of these points in the programmes’ defence, and each is true; the beneficiary’s insecurity is not a refutation of them but the price of them.

Q: What would an urban employment guarantee gain from being enacted as a statute, and what would it not gain?

It would gain the legal half of the rural guarantee’s architecture: a defined household’s claim to demand work on a dated application, a period within which work must be provided, an unemployment allowance pricing non-delivery, works conditions and a wage discipline fixed in a text a circular could not dilute, a named forum for refusal, and continuity across political cycles of the kind Kerala’s urban employment scheme, running since 2010 as an executive programme modelled on the rural Act, cannot give its workers. It would not gain the money, the staffing to open works at the speed demand requires, or an adequate wage, since the appropriation remains annual and the delivery tier remains to be built; a state that enacted the guarantee and starved it would reproduce the rural programme’s rationing in an urban setting.

Q: What happens to beneficiaries when a welfare scheme is merged into another programme?

They are carried, if they are carried, on the successor’s terms. A merger is effected by executive decision, so the successor programme’s guidelines fix the beneficiary criteria, the benefit and the administering authority, and the predecessor’s lists have no legal force of their own: households recognised under the earlier instrument must qualify afresh under the new one, and those who fall outside its criteria lose the benefit without any proceeding at which the loss could be contested. The pattern is the scheme layer’s normal metabolism, visible when the rural housing programme was restructured and renamed with a revised beneficiary list drawn from a newer census base, and when the earlier health insurance scheme for poor households was subsumed into the Ayushman Bharat cover under different criteria and a different authority.

Q: When is the scheme form the better choice for a welfare programme?

When the programme is genuinely experimental: when the beneficiary definition, the benefit level and the delivery channel are all expected to be revised as evidence arrives, when the failure modes are not yet known, and when beneficiaries are not exposed to irreversible harm while the design settles. In that situation enactment freezes a first draft into the hardest text in the legal system to amend, multiplies the forums in which early errors can be litigated, and purchases continuity for a design that has not yet earned it. The scheme form is also the honest instrument for emergency response to a sudden and event-defined need, where speed is the substance of the provision. The choice is discreditable only when flexibility is claimed for a mature programme whose design is settled and whose beneficiaries plan their survival around it.

Q: How does portability of a welfare benefit differ between a statutory entitlement and a scheme?

In the mechanics, barely at all; in the legal position, completely. Portability has been built administratively across both layers: the ration card’s national portability lets a food statute beneficiary draw grain in a state other than the state of registration, and cash benefits under schemes travel through the same bank accounts and identity-seeded databases as statutory wages. The pipe is shared, and its quality does not depend on the source of the programme it carries. The difference appears when the travelling beneficiary is refused at the destination: the statutory beneficiary carries the Act, its definition and its forum into the new state, while the scheme beneficiary carries a place on a list whose governing guideline the home administration issued and the destination administration administers as a matter of programme practice.

Q: Why does a statutory grievance officer change a beneficiary’s position in a way a scheme helpline cannot?

Because the officer’s authority comes from outside the programme being complained about. A statutory grievance officer, a District Grievance Redressal Officer under the food statute, a grievance redressal officer in an establishment under the disability statute, a Maintenance Tribunal under the senior citizens statute, holds a text-given power to inquire into a refusal and direct the administration to correct it, and the refusal itself is measured against the Act rather than against the programme’s view of its own convenience. A helpline or complaint portal is staffed by the programme it serves, can correct an error the programme concedes, and cannot entertain the complaint that matters most, that the programme has been narrowed or its criteria changed. The officer’s existence also disciplines the counter in advance, since a refusal that can be carried to a named authority is made more carefully than one that can only travel upward through the refuser’s own chain.

Q: Can a scheme beneficiary claim a legitimate expectation that the programme will continue?

Not in any form that binds the government to continue it. Indian courts protect a beneficiary’s expectation that a subsisting programme will be administered fairly and according to its announced criteria, and an individual who has been promised a benefit under a scheme’s own terms and has ordered their affairs on that promise may in a defined case hold the administration to the promise while the scheme stands. But the doctrine does not convert a policy into an entitlement: a programme may be revised prospectively or discontinued in the public interest by the authority that created it, and no expectation arising from its past generosity can compel its future existence. The beneficiary’s true security under a scheme is political, the weight of the beneficiary class, and not legal.

Q: How should a legislator test whether a proposed welfare programme needs a statute?

By running the dimensions as questions. What happens to the beneficiary if the programme is interrupted for a season: if the harm is severe and irreversible, the programme is a candidate for enactment. Can the beneficiary class protect itself politically: if it is dispersed, poor and unorganised, the remedy a statute creates substitutes for the voice it lacks. Is the design settled: if the definition, the benefit and the delivery channel are still expected to change, enactment is premature. And can the appropriation plausibly meet the entitlement at full demand: if it cannot, the legislator should either narrow the entitlement honestly in the text or accept that the statute will ration, and should write the funding share, the indexation and the identification machinery into the bill with the same care as the right itself, because those provisions decide whether the enacted promise is usable.

Q: Does converting a scheme into a statute increase its administrative cost?

It changes the cost’s character more than its size, and in both directions. A statute adds machinery a scheme can omit: statutory officers and commissions to be appointed and staffed, social audits and vigilance structures to be run, allowances and compensation to be computed, and litigation to be answered, all of which are real administrative burdens carried in the programme’s cost. Against that, a mature scheme already carries much of the same apparatus in executive form, its own officers, portals, audits and grievance channels, and the statutory version can be cheaper to defend because its criteria stop being renegotiated at every appraisal. The honest answer is that the measurable administrative difference is smaller than either side claims, and the real cost of enactment is not administrative at all: it is the loss of the power to revise quietly, which is a cost to the government and the entire point of the exercise for the beneficiary.