Indian public discussion uses one word for two different legal things. A rural household that demands work under the employment guarantee, a widow who receives a monthly pension, a child who eats a cooked meal at school, a family whose hospital bill is paid under a health cover programme: all four are described, in official speeches and in ordinary speech alike, as beneficiaries of government schemes. The description conceals the single most important fact about each benefit. Two of those four rest on Acts of Parliament that define the beneficiary, impose a duty on the state, and create machinery through which a denial can be challenged. The other two rest on executive decisions that a government may expand, narrow, rename or discontinue without asking any legislature. Welfare laws in India cannot be understood as one body of law until that distinction is drawn, and almost every practical question in the field, from who qualifies to what a refused applicant can do next, is answered by it.

This article maps the whole field on that basis. It begins with the constitutional layer, a set of directive principles that instruct the state to build a welfare order but cannot be enforced in a court. It then traces the judicial bridge built across that gap, through which the right to life was read to cover livelihood, food, shelter, health and education. It describes the older statutory models, contributory social insurance for the organised workforce and a maintenance duty placed on families, before turning to the rights turn of the 2000s, when Parliament converted a series of programmes into statutory entitlements with defined beneficiaries and grievance machinery. It sets out the scheme layer that has always existed alongside the statutes and has grown faster in the years since. It then places the two layers side by side in a single table, tests the two standard arguments made for each model, and closes with the practical method by which a reader can identify the legal source of any programme they encounter. The claim that organises everything that follows can be stated in one sentence: the source decides the remedy, because whether a benefit rests on an Act or on an executive order determines whether a person denied it holds a claim or only a complaint.

The vocabulary of the field does some of the concealing. Statutes carry the names of their subjects and the year of their enactment, and they are cited by those names in courts and offices. Schemes carry the names of their purposes or their patrons, styled as a yojana, a mission or an abhiyan, and are cited by acronym. The naming is not a reliable guide by itself, since a mission can be created to implement a statute and an Act can christen a programme, but it is a serviceable first signal: an instrument that cannot be cited by year and number, that has no existence outside the budget documents and the guidelines of a ministry, is an instrument whose promises are exactly as durable as the administrative will behind them. The sections that follow take the instruments in the order of the framework’s construction, from the constitutional instruction outward to the counter at which the citizen stands.

The constitutional foundation: instructions without a remedy

The Constitution of India does not contain a chapter of welfare rights. What it contains is Part IV, the directive principles of state policy, a set of instructions addressed to legislatures and governments that Article 37 declares to be fundamental in the governance of the country while expressly barring their enforcement by any court. The welfare framework of the Republic is built on that deliberate asymmetry: the ends are stated in the constitutional text, the means are left to legislation, and the citizen who is denied a benefit must therefore find the source of the claim in a statute or a scheme rather than in Part IV itself.

The principles that carry the welfare field can be identified precisely. Article 38 directs the state to promote the welfare of the people by securing a social order in which justice, social, economic and political, informs all institutions of national life, and to minimise inequalities in income, status, facilities and opportunities. Article 39 includes among its specified policies the direction that citizens have the right to an adequate means of livelihood. Article 41 is the centre of the field: it requires the state, within the limits of its economic capacity and development, to make effective provision for securing the right to work, the right to education and the right to public assistance in cases of unemployment, old age, sickness and disablement, and in other cases of undeserved want. Article 42 directs the state to make provision for securing just and humane conditions of work and for maternity relief. Article 43 directs it to endeavour to secure a living wage, conditions of work ensuring a decent standard of life, and the promotion of cottage industries. Article 46 requires the state to promote with special care the educational and economic interests of the weaker sections of the people, and in particular of the Scheduled Castes and the Scheduled Tribes, and to protect them from social injustice and all forms of exploitation. Article 47 makes it a primary duty of the state to raise the level of nutrition and the standard of living of its people and to improve public health.

Which directive principles does the welfare framework rest on?

The framework rests chiefly on Articles 38, 39, 41, 42, 43, 46 and 47: the welfare order, an adequate means of livelihood, work, education and public assistance in unemployment, old age, sickness and disablement, just and humane conditions of work with maternity relief, a living wage, protection of weaker sections, and the duties on nutrition and public health.

One phrase of Article 41 deserves a moment’s attention because the whole subsequent law of assistance is a gloss on it. The principle promises public assistance in cases of undeserved want, an expression inherited from the vocabulary of the poor law, in which relief was graded by the desert of the applicant. The statutes and schemes of the Republic have silently rewritten the phrase. The entitlement statutes assist on the basis of a defined condition, unemployment, age, disability, pregnancy, bereavement, without inquiring into desert; the food statute’s categories and the assistance programme’s pension classes alike distribute by status rather than by character. Whatever the faults of identification by category, and they occupy a later section of this article, the abandonment of the desert inquiry is one of the framework’s settled achievements, and it was achieved by legislation rather than by interpretation, which is where the Constitution intended such choices to be made.

Two features of this layer matter for everything built on it. The first is the qualifying language of Article 41 itself. The duty to make effective provision is expressly made subject to the limits of the state’s economic capacity and development, so the principle concedes in its own text the fiscal argument that later reappears inside the statutes, where entitlements are formally open-ended and operationally rationed by annual appropriations. The second is the institutional consequence of Article 37. Because no court may enforce a directive principle as such, a Parliament that wishes to make a welfare promise real must translate it into a statute, and the history of Indian welfare law is substantially the history of which principles received that translation, in what order, and with what machinery attached. Education received it through a constitutional amendment that moved part of the promise into the fundamental rights chapter itself. Food, work and disability received it through ordinary statutes. Old age pensions, by contrast, were never translated at the Union level at all, and they remain the largest welfare commitment in the country that rests on executive decision alone.

The placement of these duties in a non-enforceable chapter was a considered choice, not an oversight. The Constituent Assembly, following the model of the Irish Constitution, adopted the principles as instructions to future legislatures, and Dr B. R. Ambedkar defended them in the Assembly as instruments of instruction to whoever held power: a government that ignored them would answer for the disregard to the electorate rather than to a court. Proposals to make the principles judicially enforceable were considered and not adopted. The consequence is written into the structure of the welfare state that followed. Political accountability, through elections and legislatures, was made the primary sanction for welfare failure, and legal accountability was made available only where a legislature chose to create it. Every statute and every scheme in this article is a product of that choice, and the difference between the two is the difference between a field where the Assembly’s design was completed by legislation and a field where it was left to operate through politics alone.

The Supreme Court has given Part IV a constitutional weight that the non-enforceability clause does not at first suggest. In Minerva Mills v Union of India, decided in 1980, the Court held that the harmony and balance between the fundamental rights in Part III and the directive principles in Part IV is an essential feature of the basic structure of the Constitution, and that neither chapter has primacy over the other. The Forty-second Amendment of 1976 had itself enlarged the welfare content of Part IV, adding principles on free legal aid, on the participation of workers in the management of industries, and on the protection of the environment, and recasting the principle on children so that it speaks of opportunities and facilities for children to develop in a healthy manner. The principles therefore operate as the Constitution’s standing brief to the legislator: they supply the stated purpose that courts use when they interpret welfare statutes, they supply the vocabulary in which new entitlements are demanded, and they supply the standard against which the scheme layer’s claim to be a welfare policy is judged. What they never supply, by themselves, is a defendant, a duty-bearer with a name, or a forum.

The first great translation of the principles into statute was not in the field this article has still to describe but in land. The abolition of zamindari and the first generation of land reform legislation were defended as the implementation of Article 39’s direction that the ownership and control of the material resources of the community be distributed to subserve the common good and that the operation of the economic system not result in the concentration of wealth. When the courts struck elements of that programme down on fundamental rights grounds, Parliament answered with the Constitution (First Amendment) Act, 1951, inserting Articles 31A and 31B and the Ninth Schedule to place specified welfare legislation beyond the reach of Part III challenge. The episode fixed the template the field has followed since: a directive principle supplies the programme, a statute embodies it, a fundamental right is pleaded against it, and the constitutional settlement is renegotiated around the embodiment. It also taught the lesson the rights turn relearned, that a principle unembodied remains a debating point, while a principle embodied in an Act acquires beneficiaries, officers and a litigation life of its own.

The distinction between the two constitutional chapters also explains a persistent error in public writing about this field, which is to describe the directive principles as enforceable, or to describe a scheme as the fulfilment of a constitutional right. Neither is accurate. A directive principle is a direction to the lawmaker. A scheme is an act of the executive. A right, in the sense that matters to a person who has been refused something, exists only where a legislature has created one in a form a court or a statutory officer can apply, or where a court has located the substance of a principle inside an enforceable fundamental right. That second route is the subject of the next section.

The judicial bridge: Article 21 and the monitored entitlement

The bridge between Part IV and an enforceable claim was built by the Supreme Court through Article 21, the guarantee that no person shall be deprived of life or personal liberty except according to procedure established by law. Across a series of decisions, the Court read the word life to mean more than animal existence, and in doing so it carried several directive principles into the enforceable chapter without amending the constitutional text.

The propositions that matter for welfare law are specific and can be stated narrowly. In Olga Tellis v Bombay Municipal Corporation, decided in 1985, the Court treated livelihood as within the content of the right to life, reasoning that a person deprived of the means of living is deprived of life itself, in a case brought by pavement and slum dwellers facing eviction in Bombay. In Shantistar Builders v Narayan Khimlal Totame, decided in 1990, the Court counted shelter among the basic needs comprehended by the right to life. In Unnikrishnan v State of Andhra Pradesh, decided in 1993, the Court held that children up to the age of fourteen had a right to free education flowing from Article 21 read with the directive principles, a holding that prepared the ground for the Constitution (Eighty-sixth Amendment) Act, 2002, which inserted Article 21A and made free and compulsory education for children between six and fourteen a fundamental right in terms. In Paschim Banga Khet Mazdoor Samity v State of West Bengal, decided in 1996, the Court held that the failure of government hospitals to provide timely emergency medical treatment violated Article 21, and it connected that conclusion to the Article 47 duty regarding public health.

The technique was not confined to the subsistence cases. In Randhir Singh v Union of India, decided in 1982, the Court read the directive principle of equal pay for equal work in Article 39(d) into the equality guarantees and applied it as an enforceable claim for drivers in a public establishment, an early demonstration that the principles could supply content to the fundamental rights even in the law of employment, the field where welfare and labour law meet. The pattern across these decisions is consistent enough to state as a rule of the framework: a directive principle becomes litigable when a court is persuaded that its substance is already inside the right to life or the equality code, and the persuasion succeeds most readily where the claim concerns a minimum, a floor of food, treatment, shelter or schooling below which life with dignity is impossible to describe. Claims above the floor, to a quantum, a rate or a standard of provision, have been left to legislation, which is why the judicial bridge, for all its traffic, never replaced the need for the statutes that form the centre of this article.

The most consequential welfare litigation of the period was not a single judgment but a continuing proceeding. In People’s Union for Civil Liberties v Union of India, begun in 2001 against the background of drought and accumulated food stocks, the Supreme Court issued a long sequence of interim orders that treated the beneficiaries of the government’s own food and nutrition programmes as holders of entitlements under Article 21, directed that cooked midday meals be served in government and government-assisted primary schools rather than dry rations, and supervised compliance through commissioners appointed by the Court. The technique, known as continuing mandamus, kept the proceeding open while the Court monitored delivery over many years. Its importance for this field is structural. Before Parliament enacted a food security statute, the Court had already demonstrated that the entitlements contained in executive programmes could be policed as legal obligations when they were anchored to the right to life, and the statute that arrived in 2013 legislated for a field whose expectations the litigation had formed.

Two further decisions show the bridge carrying a different kind of traffic. In Municipal Council, Ratlam v Vardhichan, decided in 1980, the Supreme Court upheld an order directing a municipality to discharge its statutory sanitation duties for the benefit of the residents of a locality who had no effective remedy against the nuisance the neglect had created; the case stands for the proposition that a welfare duty written into a statute, even a municipal statute of the most routine kind, is a legal obligation a court will enforce against the body that bears it, and that financial incapacity is not a defence a public body can plead against its own statute. In Bandhua Mukti Morcha v Union of India, decided in 1984, the Court treated the release and rehabilitation of bonded labourers as a requirement of Article 21 read with the prohibition of forced labour in Article 23, and supervised the identification and rehabilitation process through commissions. The two decisions mark the bridge’s real function. Where a duty exists in some legal text, whether a municipal Act, a labour statute or the Constitution itself, the courts have been willing to construct supervision around it. Where no duty exists in any text, there is nothing for supervision to attach to, and that is the situation of the pure scheme beneficiary.

The scheme beneficiary is not wholly outside this jurisprudence, and accuracy requires the residue to be stated. In Ramana Dayaram Shetty v International Airport Authority of India, decided in 1979, the Supreme Court held that the state, in distributing its largesse, does not act as a private person may: its action must satisfy the test of reasonableness and public interest, and arbitrariness in the distribution of benefits is reviewable under the equality guarantee of Article 14. The doctrine gives the beneficiary of a subsisting scheme a real if limited protection. A guideline must be applied as written, an exclusion that the guideline does not authorise can be struck down, and a benefit distributed by whim or favour can be recalled to the programme’s own criteria. What the doctrine cannot do is equally definite. It cannot convert the guideline into a statute, prevent its prospective revision, or oblige the state to continue a programme it has lawfully discontinued. The protection attaches to the administration of the scheme, not to its existence, and the distance between those two objects is the distance between the layers that this article exists to measure.

The limits of the judicial bridge are as instructive as its reach. Article 21 litigation can establish that a class of provision may not fall below a floor of dignity, and it can compel a government to operate its own programmes honestly. It is poorly suited to fixing the content of a national entitlement: the rates, the coverage percentages, the identification criteria and the funding shares that make a welfare promise administrable are legislative choices, and the courts have generally stopped short of writing them. The bridge therefore produced pressure and precedent rather than a code. The code, when it came, was written by Parliament in the decade that followed, and the statutes it produced are the centre of the field.

Before the Constitution: where the two layers come from

The division between statutory welfare and executive welfare is older than the Republic. The Karachi Resolution of 1931, in which the Indian National Congress set out its programme of fundamental rights and economic and social change, promised a living wage for industrial workers, protection of women workers and a prohibition of child labour in factories, and the Sapru Committee of 1945 proposed that the coming constitution divide its rights into justiciable and non-justiciable classes, the division the Constituent Assembly eventually enacted as Parts III and IV. Under the colonial administration, relief of distress had been a departmental activity governed by famine codes and executive instructions, while the first statutory interventions came in the employer-liability and insurance forms: a compensation statute for employment injuries in 1923 and, following the report of the Royal Commission on Labour of 1931, a gradual construction of the insurance model that the Acts of 1948 and 1952 completed. The pattern visible in 1950 was therefore already the pattern of the present field: a small statutory core addressed to the organised industrial worker, and a large discretionary penumbra in which relief depended on administrative decision. The Constitution supplied the penumbra with a programme, in the directive principles, and the Republic’s welfare history since then is the story of the statutory core expanding, contingency by contingency, into territory the programme had marked out.

The older statutory models: insurance contributions and family duty

Long before the rights turn, Parliament had legislated for welfare in two forms that do not fit the entitlement pattern, and both forms survive inside the present framework. The first is contributory social insurance. The Employees’ State Insurance Act, 1948 (Act 34 of 1948) created a fund financed by contributions from employers and employees, out of which insured employees receive medical care and cash benefits in sickness, maternity, disablement and dependants’ contingencies. The Employees’ Provident Funds and Miscellaneous Provisions Act, 1952 (Act 19 of 1952) applied the same financing logic to retirement savings. Under this model the benefit is neither a gift of the state nor a tax-funded entitlement of citizenship. It is an earned claim on a fund to which the worker and the employer have contributed, administered by statutory corporations, and its boundary is the boundary of the covered workforce: establishments above a threshold size and employees below a wage ceiling. The vast unorganised workforce, having no employer of record to contribute alongside it, stood outside the model altogether, a structural exclusion that later statutes and codes have repeatedly tried and repeatedly failed fully to close.

A family of statutes standing beside the insurance model places the cost on the employer directly rather than on a fund. The compensation statute of 1923 (Act 8 of 1923), renamed the Employees’ Compensation Act, makes the employer liable to pay compensation for personal injury caused by accident arising out of and in the course of employment, on scales the Act prescribes, without proof of negligence. The Maternity Benefit Act, 1961 (Act 53 of 1961) obliges the employer to pay maternity benefit at the average daily wage for the statutory period of absence, a duty whose duration the amendment of 2017 extended to twenty-six weeks for a woman with fewer than two surviving children, as the series’ labour cluster records. The Payment of Gratuity Act, 1972 (Act 39 of 1972) obliges the employer to pay a terminal gratuity on superannuation, retirement, resignation after the qualifying service, or death or disablement. These statutes are welfare law in the strictest sense, enforceable claims arising on defined events, but their debtor is a private employer and their reach is again the reach of the formal employment relation. Between them, the insurance Acts and the employer-liability Acts constructed a complete statutory welfare state for a minority of the workforce, and the completeness of the construction for that minority is part of the explanation for the lateness of provision for everyone else: so long as welfare legislation meant labour legislation, the person without an employer had no statutory address.

The second older form places the welfare duty on the family rather than on the state. It has its own statutory lineage: the personal laws carried maintenance obligations all along, the Hindu Adoptions and Maintenance Act, 1956 (Act 78 of 1956) obliging a Hindu to maintain aged or infirm parents unable to maintain themselves, and the summary maintenance jurisdiction of the criminal procedure codes, carried from the Code of 1973 into its successor code, empowering a magistrate to order a monthly allowance for parents who cannot maintain themselves, on a religion-neutral footing. Its fullest statutory expression belongs to the second wave described below, the Maintenance and Welfare of Parents and Senior Citizens Act, 2007, but the idea is older than any single Act: that the first obligation to maintain an aged parent or an indigent relative rests on children and heirs, with the state supplying a tribunal to enforce the obligation and a residual institution, the old age home, for those whom the family model cannot reach. Between insurance for the organised worker and maintenance enforced against the family, the older framework left the largest group of all, the unorganised worker without contributory coverage and the destitute person without family means, to discretionary relief. That uncovered space is where the modern scheme layer grew, and it is also the space the rights turn legislated for first.

The third older model: welfare boards financed by a cess

Between the insurance model and the rights turn, Indian welfare law developed a third financing device that fits neither the contributory corporation nor the tax-funded entitlement: the occupation-wise welfare board financed by a cess on production or construction. Its fullest expression is the pair of statutes of 1996 for construction workers, the Building and Other Construction Workers (Regulation of Employment and Conditions of Service) Act, 1996 (Act 27 of 1996) and the Building and Other Construction Workers’ Welfare Cess Act, 1996 (Act 28 of 1996). The second Act levies a cess on the cost of construction, the proceeds are credited to state welfare boards constituted under the first Act, and a construction worker registered with a board as a beneficiary draws benefits from the fund: assistance in case of accident, pension on reaching sixty, assistance for children’s education and medical expenses, and other measures the board frames. Kerala had built occupation-wise welfare funds by state legislation over earlier decades, for toddy workers, headload workers and other trades, financed by combinations of employer, worker and state contribution, and the construction model generalises that design.

The board model matters to the field map because it relocates every element of the entitlement question. The beneficiary is defined, but by registration with a board rather than by residence or household status, and the history of the construction cess is the standard demonstration of the identification problem in its fiscal form: cess collections accumulated in state funds on a large scale while registration of the workers entitled to draw on them lagged far behind, a gap the Supreme Court examined in proceedings on the implementation of the 1996 Acts. The duty-bearer is a board, not a department. The money is a hypothecated cess, not a budget line and not an insurance contribution shared with an employer of record. And the remedy for a refused benefit runs to the board that holds the fund. A reader applying the source test to a board benefit will find it genuinely intermediate: statutory in origin, fund-based in finance, and administrative in delivery, which is precisely why the model was attractive for the informal workforce that the insurance model could not reach and the entitlement model had not yet covered.

The rights turn, first half: work and food

The statutory sequence that defines the modern field begins with employment, and it begins after a long apprenticeship in the scheme form. Rural public works as relief are as old as the famine codes, and independent India ran a succession of executive employment programmes: the National Rural Employment Programme of 1980, the Rural Landless Employment Guarantee Programme of 1983, the Jawahar Rozgar Yojana of 1989, the Employment Assurance Scheme of 1993 and the Sampoorna Grameen Rozgar Yojana of 2001, among others. Each was a budgetary programme with guidelines, targets expressed as person-days to be generated, and no individual entitlement: a worker could not demand work, only hope that a sanctioned work was open in the village. Maharashtra’s employment guarantee, by contrast, rested on a state statute, the Maharashtra Employment Guarantee Act of 1977, and its two decades of operation supplied the rights turn with its working proof that a statutory guarantee could be administered at scale, a proof the national debate drew on directly. The national Act of 2005 is best read as the generalisation of the Maharashtra experiment and the repudiation of the programme sequence that preceded it: Parliament looked at forty years of schemes and enacted the thing the schemes had each stopped short of being. The National Rural Employment Guarantee Act, 2005 (Act 42 of 2005), which received assent on 5 September 2005, came into force on 2 February 2006 in two hundred districts, and was extended to all rural districts from 1 April 2008, guarantees one hundred days of wage employment in a financial year to every rural household whose adult members volunteer to do unskilled manual work. It was renamed the Mahatma Gandhi National Rural Employment Guarantee Act in 2009. Its drafting broke with the scheme pattern in four deliberate ways. The beneficiary is defined by the statute itself rather than by a guideline. The duty is placed on the state government in mandatory terms, and an application for work triggers a dated receipt and a fifteen-day window within which work must be provided. Section 7 of the Act provides that where work is not provided within that window, the applicant is entitled to a daily unemployment allowance at rates the Act fixes by reference to the wage rate, so that non-delivery itself generates a compensatory claim. Section 17 requires the Gram Sabha to conduct regular social audits of the works, building a public audit into the delivery structure. Section 6 empowers the Union to notify a wage rate for the programme distinct from the agricultural minimum wage of the state, a provision that has kept the programme’s wage a subject of continuing federal negotiation, and the Act permits up to fifty additional days of employment in a financial year in areas affected by drought or natural calamity when so notified, so that the hundred-day floor is a floor and not a ceiling in the conditions of greatest need. The funding pattern is written into Section 22: the Union bears the entire cost of wages for unskilled manual work and three-fourths of the material cost, while the state bears the remainder and the unemployment allowance. The architecture of the Act, its Programme Officers, District Programme Coordinators and panchayat-level execution, is examined in the series’ treatment of the right to work under the employment guarantee statute, which owns the statute’s detail; what matters for the field map is that in 2005 a demand-driven programme crossed the line into statutory entitlement, and every later welfare statute was drafted in awareness of that crossing.

Food followed eight years later, out of an equally long scheme apprenticeship. Public distribution of foodgrains began as wartime rationing, continued as a general executive system, and was converted in 1997 into the Targeted Public Distribution System, under which the population was divided by administrative criteria into households above and below the poverty line with different entitlements to grain at different prices. The targeting was a guideline decision, revisable at will, and the litigation of 2001 onward exposed its exclusions to judicial scrutiny years before Parliament acted. When the statute came, it therefore did something more specific than create an entitlement: it froze the coverage question at the percentages the Act states, displaced the poverty line as the operative legal boundary, and subjected the identification that remained to states to the discipline of a statutory text. The National Food Security Act, 2013 (Act 20 of 2013) received assent on 10 September 2013 and was deemed to have come into force on 5 July 2013. It converted the targeted public distribution system from an executive programme into a statutory entitlement for up to seventy-five per cent of the rural population and fifty per cent of the urban population, percentages the Act itself fixes. Section 3 entitles persons belonging to priority households to receive five kilograms of foodgrains per person per month at the subsidised prices in Schedule I of the Act, three rupees per kilogram for rice, two rupees for wheat and one rupee for coarse grains, and entitles households covered under the Antyodaya Anna Yojana to thirty-five kilograms per household per month. The Act reaches beyond grain. Section 4 entitles pregnant women and lactating mothers to a maternity benefit of not less than six thousand rupees, a figure stated in the statute itself. Section 5 entitles children to nutritional support meeting the standards of Schedule II, delivered through meals at anganwadis and midday meals in schools, so that two programmes born as executive schemes were given a statutory floor. Section 8 provides for a food security allowance, payable by the state government where the entitled quantities are not supplied. Section 29 requires vigilance committees at the state, district, block and fair price shop levels to supervise implementation, and the Act’s transparency provisions require records of the distribution system to be placed in the public domain and social audits to be conducted, carrying the employment statute’s audit idea into the food system. The identification of eligible households is left to the state governments within criteria the Act frames, an allocation of power that the field’s later disputes have repeatedly returned to. The full working of the entitlement, its grievance officers and its State Food Commissions is set out in the national food security guide, which owns the statute; its place in this article is as the second and largest pillar of the rights turn, and as the clearest proof that the distinction between statute and scheme is not academic, because the same ration shop transaction is a claim under one description and a favour under the other.

The statute also contains its own reform agenda, which distinguishes it from a simple codification of the system it replaced. Its provisions on the reform of the targeted public distribution system direct attention to the doorstep delivery of foodgrains, the use of information technology to identify duplicates and false cards, the diversification of commodities distributed, and the leveraging of the banking and postal networks, and they contemplate the introduction of schemes of cash transfer or food coupons in a manner that protects the entitlement while the mechanism is tested. The presence of a reform chapter inside an entitlement statute is instructive: Parliament in 2013 did not suppose that the form of delivery was settled by the creation of the right, and the subsequent history of the food system, digitised ration cards, electronic point of sale devices at fair price shops and national portability, has been a working out of the statute’s own reform directions by administrative means. Right, machinery and reform programme coexist in the one text, which is the mature form of the rights-turn statute and the reason the food Act repays closer reading than its reputation as a subsidy law suggests.

Two features of the food statute deserve emphasis here because they qualify the general claim rather than merely illustrate it. The entitlement is defined against population percentages, which means the Act guarantees a share of coverage rather than coverage of every person in need, and the identification of the households inside the share is a state function whose quality decides who actually holds the right. And the statute’s own text contemplates reform of the delivery system, including the possibility of cash transfers in place of grain, so the form of the entitlement is more open than its existence. A statutory right, in other words, settles the question whether a benefit may be denied arbitrarily. It does not by itself settle the questions of identification, adequacy or delivery, and the food statute’s record is the field’s standard evidence for that proposition.

The employment statute’s content is completed by its Schedules, which the field’s debates often overlook. Schedule I lists the permissible works, concentrated on water conservation, drought proofing, irrigation, renovation of water bodies, land development, flood control and rural connectivity, so that the wage is statutorily tied to the creation of a defined class of rural assets. Schedule II sets the conditions of work: the wage and material ratio to be maintained, the worksite facilities of drinking water, rest and first aid and childcare where young children are present at the site, the prohibition of contractors and of labour-displacing machinery, and the requirement that at least one-third of the beneficiaries to whom work is provided be women. The Schedules matter to the classification exercise because they show how much of the programme’s character Parliament chose to fix in law rather than leave to the operational guidelines that every scheme lives by: under the employment statute, a ministry circular cannot lawfully convert the programme into a contractor-driven works scheme or strip the worksite facilities, because the text forbids it. The entitlement’s durability lies as much in these unglamorous conditions as in the hundred days themselves.

The rights turn, second half: education and forest livelihoods

Education took a different constitutional route to a similar destination, and the route began with a missed deadline. Article 45, as enacted in 1950, directed the state to endeavour to provide free and compulsory education for all children until the age of fourteen within ten years of the Constitution’s commencement. The decade passed without the provision being made, and the principle stood for forty years as the standard exhibit of the directive principles’ unenforceability, until the judicial route and then the Eighty-sixth Amendment converted its substance into a fundamental right and the 2009 statute into an administrative code. No other item in the welfare framework displays the full life cycle so completely: principle, default, litigation, constitutional amendment, statute, and then the long administrative labour of making the statute true. The Right of Children to Free and Compulsory Education Act, 2009 (Act 35 of 2009), in force from 1 April 2010, gave statutory content to Article 21A by guaranteeing free and compulsory elementary education to every child between the ages of six and fourteen, imposing duties on governments to establish neighbourhood schools, on schools to meet prescribed norms, and on private unaided schools to admit children from weaker sections and disadvantaged groups to at least twenty-five per cent of the entry-level class, with reimbursement from the state. The statute’s grievance route runs through the local authority, and its monitoring through the commissions for protection of child rights at the Union and state levels. Its central admissions provision survived constitutional challenge in Society for Un-aided Private Schools of Rajasthan v Union of India, decided in 2012, where the Supreme Court upheld the twenty-five per cent obligation for non-minority unaided schools as a reasonable measure in aid of the fundamental right, and the judgment in Pramati Educational and Cultural Trust v Union of India, decided in 2014, held the Act inapplicable to minority institutions, aided and unaided alike, on the strength of the minority rights in Article 30. The pair of decisions illustrates the education statute’s distinctive position in the framework: alone among the rights-turn statutes, it implements a fundamental right directly, and its duties are therefore tested against the competing fundamental rights of other constitutional actors rather than against appropriations alone. Where the employment and food statutes created entitlements to a commodity or a wage, the education statute created an entitlement to a service delivered by a mixed system of public and private providers, and its enforcement record shows the strain that follows when a statutory duty is laid on institutions the state does not directly run. It belongs in the welfare framework because its function is welfare in the Article 41 and Article 45 sense, the conversion of a directive principle first into a judicially recognised right, then into a fundamental right, and finally into an administrative code.

The forest rights statute belongs to the framework by a different logic again, because what it distributes is not expenditure but recognition. The Scheduled Tribes and Other Traditional Forest Dwellers (Recognition of Forest Rights) Act, 2006, which is Act 2 of 2007 notwithstanding the year in its short title, in force from 31 December 2007, recognises and vests forest rights in forest dwelling Scheduled Tribes and other traditional forest dwellers: rights to hold and live on forest land under occupation, to own and dispose of minor forest produce, to community rights of use and access, and to the conversion of forest villages into revenue villages. Its machinery is deliberative rather than bureaucratic: the Gram Sabha initiates the determination of claims, a Sub-Divisional Level Committee examines them, and a District Level Committee takes the final decision, with petitions available within the committee structure itself. For the field map, the statute shows that a welfare entitlement need not be a payment. The security it creates is security of tenure and livelihood against the state itself, and the denial it guards against is not the refusal of a benefit but eviction and the denial of access. Its presence in the same framework as the pension and the ration is a reminder that the statutory layer of Indian welfare law was built by several movements with different objects, which happen to share a legal form.

The second statutory wave: disability, ageing, identity and the street

The statutes that followed the rights turn extended its form to fields the first wave had not reached, though each adapted the form to a different kind of claim. The sequencing followed the availability of the same three conditions that had produced the first wave, in updated forms. Disability legislation was driven by an international instrument, the United Nations Convention on the Rights of Persons with Disabilities, which India ratified in 2007 and which required the replacement of the welfare-and-charity model of the 1995 statute with a rights model; the senior citizens statute followed a national policy on older persons adopted by executive decision in 1999, an instance of the migration this article tracks, in which a policy document rehearses a commitment for years before a legislature converts it into obligation; the transgender statute followed a judgment. Constituency, text and demonstration, the conditions identified in the politics of the first wave, recur because they are the conditions on which legislatures in this field act. The Rights of Persons with Disabilities Act, 2016 (Act 49 of 2016), which received assent on 27 December 2016 and replaced the statute of 1995, is principally a rights and non-discrimination code rather than a benefits statute: it recognises twenty-one categories of disability, defines the person with benchmark disability by a threshold of forty per cent, reserves not less than five per cent of seats in government and government-aided higher educational institutions and not less than four per cent of vacancies in government establishments for persons with benchmark disabilities, obliges establishments to maintain grievance redressal officers for disability complaints, and creates Chief and State Commissioners with powers to inquire into deprivation of rights. Its accessibility provisions impose duties of a different texture again: standards for the built environment, transport and information and communication that public buildings and services must meet, with existing public buildings required to be made accessible within the period the Act prescribes, so that the entitlement here is to a physical and informational environment, enforceable against governments and, in the respects the Act specifies, against private establishments as well. Its welfare content is therefore chiefly about enforceable equality of access, backed by named officers and commissioners. One provision deserves notice for what it reveals about the whole framework: Section 24 of the Act directs the appropriate government to formulate schemes and programmes for the social security of persons with disabilities, and it does so in the Constitution’s own qualifying words, within the limits of its economic capacity and development. The directive principle’s escape clause, drafted in 1949 for a non-enforceable instruction, is reproduced inside an enforceable statute, where it converts that part of the Act back into a duty to design programmes rather than a schedule of benefits. The two layers of the framework are thus not merely neighbours; the older layer survives inside the newer one, and a careful reader of any welfare statute will look for the qualifying phrase that marks the boundary between what the Act entitles and what it merely encourages. The series’ pillar treatment of the Rights of Persons with Disabilities Act carries its provisions in full. For this map, its significance is that the entitlement form proved capable of carrying claims that are not about money at all.

The Maintenance and Welfare of Parents and Senior Citizens Act, 2007 (Act 56 of 2007), which received assent on 29 December 2007, took the opposite course on the question of who owes the duty. Instead of a state-funded pension, it makes the maintenance of parents and senior citizens who cannot maintain themselves a legal obligation of their children and, in defined cases, of relatives who inherit or possess their property. It creates Maintenance Tribunals, each presided over by an officer of the rank the Act specifies from the state’s civil administration, to decide applications summarily and Appellate Tribunals above them, empowers a tribunal to void a property transfer made on a condition of care that the transferee then failed to honour, makes abandonment of a senior citizen an offence, and directs state governments toward the establishment of old age homes and medical support for the indigent. The entitlement is real and adjudicable, but the debtor is the family, and the state’s role is to supply the forum. The detailed working of the tribunals is treated in the article on the senior citizens maintenance statute. Read alongside the pension schemes described below, this Act is the sharpest illustration in the framework that a statutory welfare right and a state welfare payment are different instruments: old age support in India is carried simultaneously by a statute that obliges children, a set of executive schemes that pay pensions, and state statutes and schemes that vary the combination.

The Transgender Persons (Protection of Rights) Act, 2019 (Act 40 of 2019), which received assent on 5 December 2019 and came into force on 10 January 2020, followed the judgment in National Legal Services Authority v Union of India, decided in 2014, which had recognised transgender persons as a third gender and directed welfare measures for them. The Act prohibits discrimination against transgender persons across education, employment, healthcare and access to goods and services, provides for a certificate of identity issued by the District Magistrate on application, obliges governments to formulate welfare schemes and programmes for the community, and constitutes a National Council for Transgender Persons to advise on policy, monitor the impact of measures and redress grievances. The statute’s welfare chapter is cast in terms of governmental obligation to design schemes rather than individual entitlement to defined benefits, which places it between the two layers of this article’s map: a statute whose operative welfare content is a duty to run programmes, whose beneficiaries then depend on the programmes the duty produces. The drafting choice is contested, and the register of this cluster requires both positions in their proponents’ terms. The National Legal Services Authority judgment had directed that transgender persons be treated as socially and educationally backward classes for the purpose of reservation in education and public employment. The Act as enacted contains no reservation provision. Organisations of transgender persons and commentators supporting them describe the omission as a dilution of the judgment and object to the certificate mechanism as a bureaucratic gate placed on an identity the judgment had treated as self-determined. Defenders of the enacted text respond that the Act’s anti-discrimination provisions, its welfare chapter and the National Council supply the operative protections, that reservation is a matter for the reservation framework’s own provisions, and that a certificate is the ordinary precondition of administering any welfare benefit to a defined class. The article records the disagreement because it is the field’s clearest instance of a general phenomenon: a judicially declared position, a statute that implements part of it, and a residue that returns to the scheme layer, where its fulfilment depends on the programmes governments choose to frame.

The Street Vendors (Protection of Livelihood and Regulation of Street Vending) Act, 2014 (Act 7 of 2014) protects a livelihood rather than paying a benefit. Built on the proposition, traceable to the Article 21 livelihood cases, that street vending is a lawful occupation to be regulated rather than a nuisance to be cleared, it requires Town Vending Committees, on which elected street vendors hold not less than forty per cent of the seats, to survey vendors at intervals of five years and issue certificates of vending, forbids eviction of a vendor in the situations the Act specifies, and creates, in Section 20, committees chaired by a person who has been a civil judge or a judicial magistrate for the redressal of grievances and resolution of disputes of street vendors, with an appeal to the local authority. Its grievance design is unusual in the framework because it borrows a judicial officer for a livelihood dispute, and its beneficiary definition is occupational: the Act protects a person because of the work they do, in the place they do it. The protection against eviction is the operative right: until the survey the Act requires has been completed and certificates issued, an existing vendor is not to be evicted or relocated, and relocation thereafter must follow the Act’s own process and criteria, so that the statute converts what had been a recurring municipal drive, conducted as pure executive action, into a regulated procedure with a record and an appeal.

A neighbouring statute of the same period deserves a place on the map because it states a service entitlement in nearly pure form. The Mental Healthcare Act, 2017 (Act 10 of 2017), whose provisions the series treats in its own article, confers on every person a right to access mental healthcare and treatment from services run or funded by the government, and obliges the government to make provision for a range of community and institutional services, with the right framed for the person rather than for a beneficiary class and supported by review boards with a complaints function. It is not ordinarily counted among the welfare statutes, and its delivery machinery is the health system rather than a welfare department, but its structure is the framework’s structure: a defined right-holder, a duty on the state, a standard of service the Act describes, and a forum. Its presence at the edge of the map is a reminder that the entitlement form has been spreading by subject rather than by design, one constituency and one campaign at a time.

Social security and the labour codes: the unfinished consolidation

The exclusion of the unorganised sector from the insurance model was addressed legislatively in the Unorganised Workers’ Social Security Act, 2008 (Act 33 of 2008), which defined the unorganised worker and empowered governments to formulate welfare schemes for life and disability cover, health and maternity benefits, and old age protection, administered through boards at the Union and state levels. In structure, that Act is a statute that authorises schemes: the entitlement layer is thin, and the substance sits in the schemes framed under it. The consolidation of labour law into four codes carried the same approach further. The Code on Social Security, 2020 (Act 36 of 2020) subsumed the insurance and provident fund statutes along with the unorganised workers statute, defined the gig worker and the platform worker as statutory categories for the first time, provided for schemes for unorganised, gig and platform workers to be framed and funded through a social security fund with contributions that may include aggregator contributions, and retained registration as the gateway to benefits. The four labour codes, including the Code on Social Security, were brought into force on 21 November 2025, as the series’ labour cluster records. The Code’s definitions repay attention because they show a legislature legislating for work it cannot fit into the employment relation: a gig worker is defined as a person who performs work or participates in a work arrangement and earns from such activities outside of a traditional employer and employee relationship, and a platform worker by reference to a work arrangement through an online platform. The definitions create the beneficiary classes; the benefits await the schemes. Until a scheme is framed, funded and opened for registration, the newest categories in Indian welfare law hold a statutory identity and an inchoate claim, which is the hybrid condition in its purest form. For the welfare map, the codes matter because they confirm the durability of the hybrid: even a consolidating statute, in this field, characteristically creates the power to run schemes and a fund to run them from, rather than a schedule of individual entitlements, so the boundary between the statutory layer and the scheme layer runs through the middle of the newest statute in the field and not merely between old and new instruments.

Pensions: every model inside one contingency

Old age income security is the contingency in which the framework’s models coexist most completely, and a reader who can classify the pension instruments can classify anything in the field. The organised worker’s pension is insurance: the Employees’ Pension Scheme of 1995 is, despite its weight in workers’ lives, a scheme in the technical sense, framed under the Employees’ Provident Funds Act of 1952, financed by diverting a share of the employer’s provident fund contribution and a subvention from the Union, and paying a pension computed by a formula the scheme prescribes. The government servant’s pension was, for the older service, a defined benefit charged on the Consolidated Fund and governed by statutory rules made under the Constitution’s provisions for public services; for entrants to central civil service from 2004, the defined benefit was replaced by the National Pension System, a contributory, market-linked arrangement begun by executive decision, which acquired a statutory regulator when the Pension Fund Regulatory and Development Authority Act was enacted in 2013 (Act 23 of 2013), without the benefit itself becoming a statutory entitlement to any amount. The unorganised worker’s pension is the assistance programme described in the next section, a budgetary payment whose central share is fixed by guideline. The assured pension marketed to the same worker under the Atal Pension Yojana is a scheme product with a guaranteed pension band stated in its guidelines and a Union co-contribution for early joiners, alterable by the authority that issued it. And beneath all of these lies the family-duty statute of 2007, which assumes the first pension provider is the earner’s own children.

The lesson of the comparison is that the word pension carries no legal information at all. Two pensioners drawing the same monthly amount may stand in entirely different legal positions, one with a formula claim on a fund, one with a rule-based claim on the Consolidated Fund, one with a guideline expectation that the next appraisal may revise, and one with a tribunal order against a son or daughter. Any serious discussion of pension reform in India is, whether it knows it or not, a discussion about moving populations between these positions, and the political difficulty of pension reform is precisely that each position has occupants who experience a transfer between models as a gain or loss of legal security, not merely of money.

Health: the framework’s thinnest statute book

Health exposes the opposite condition: a field of the first importance to welfare in which the statutory layer is at its thinnest. Article 47 makes the improvement of public health a primary duty of the state, and the Article 21 cases made emergency treatment and the obligations of government hospitals litigable at the floor. But India has no general health entitlement statute: no Act defines a package of care to which every resident is entitled as the food Act defines grain or the education Act defines schooling. What exists instead is a regulatory statute book, of which the Clinical Establishments Act of 2010 is the Union’s principal registration and standards instrument for the states that adopt it, an insurance model delivering medical care to the organised workforce through the Employees’ State Insurance Corporation’s own hospitals and dispensaries, a public hospital system that rests on state administration and budget rather than on any statutory duty to treat a defined beneficiary, and, occupying the space where an entitlement statute might have stood, the largest health scheme in the framework, the Ayushman Bharat health cover discussed below. The consequence is visible in the table: the health row of Indian welfare sits almost wholly in the scheme half, and the recurring proposal for a statutory right to health, carried into legislation in Rajasthan by a statute passed in 2023 on a bill first introduced in 2022, is best understood as an attempt to move the field’s largest gap across the line this article draws. The public hospital system itself is delivered in the mission form: the National Health Mission, an executive umbrella under which the rural and urban health missions run, channels Union funds to state health societies against approved plans, staffing the sub-centres, primary health centres and community health centres through which most public care reaches its users. The mission form gives the Union a lever over a State List subject without legislating on it, because the money carries the conditions, and it gives the beneficiary no corresponding lever at all: the staffing norm of a health centre, the presence of a doctor, the availability of a drug are programme commitments, enforceable through the administrative chain and through the Article 21 floor in the extreme case, and through no forum the mission itself creates. Health is thus the field in which the framework’s oldest arrangement persists almost undisturbed: a constitutional duty, a court-policed minimum, and a vast discretionary system in between.

For the child, health and nutrition entitlements arrive by other doors: the anganwadi’s services under an executive programme, the school meal under the food statute, and schooling itself under the education statute, so that the child’s welfare position is a composite built from three instruments of three different legal kinds, a fact anyone administering or claiming for a child has to hold together without help from any single text.

The scheme layer: provision by executive decision

Beneath and beside the statutes lies the layer that most welfare spending actually passes through: programmes created by executive decision, financed through annual budgetary allocations, governed by guidelines issued by the administering ministry, and terminable or alterable by the same authority that created them. The National Social Assistance Programme, launched in 1995 as a centrally sponsored programme, is the senior instrument of the layer: it pays old age, widow and disability pensions and a family benefit on the death of a breadwinner to households below the poverty line, under criteria fixed by its guidelines. Under those guidelines the Union’s contribution to the old age pension has been set at two hundred rupees a month for a pensioner between sixty and seventy-nine years of age and five hundred rupees a month at eighty years and above, with state governments adding amounts of their own that vary widely, so that the pension a person receives is a composite of a central guideline, a state top-up and a budget line, and no part of it is stated in any statute. The Integrated Child Development Services, begun in 1975, delivers nutrition, health and early education services through anganwadi centres on the same executive basis, although the food security statute has since placed a statutory floor under the nutritional entitlements of the young child and the pregnant or lactating mother whom the anganwadi serves, so that here the two layers overlap in a single institution. The housing programmes run by the Union and the states, of which the Pradhan Mantri Awas Yojana is the principal Union instrument, provide assistance toward the construction or upgrading of houses on guideline criteria. The Pradhan Mantri Kisan Samman Nidhi, begun in 2019, pays income support of six thousand rupees a year in three instalments to landholding farmer families, a figure fixed by the scheme’s guidelines and alterable by the same route.

Two components of the assistance programme show the layers interlocking. The Annapurna component, added in 2000, supplies ten kilograms of foodgrains free of cost each month to senior citizens who are eligible for the old age pension but not receiving it, an executive bridge across a gap in an executive programme. The National Family Benefit Scheme pays a lump sum, fixed by the guidelines at twenty thousand rupees, to a below poverty line household on the death of its primary breadwinner. Neither component has any statutory text; both are revised, when they are revised, by the ministry that administers them.

The maternity provisions show the interlock in its most instructive form, because the same contingency is covered by both layers at once and the two covers differ. The food security statute, in Section 4, entitles a pregnant woman and lactating mother to a maternity benefit of not less than six thousand rupees: the floor is in the Act, and it binds. The Pradhan Mantri Matru Vandana Yojana, the maternity benefit scheme administered by the Ministry of Women and Child Development, pays five thousand rupees for the first living child, under conditions its guidelines prescribe, including registration and institutional delivery conditions. The scheme payment is not the statutory benefit under another name, and treating the two as interchangeable, as public discussion routinely does, misstates both: the statutory floor is higher than the scheme amount, attaches to the wider class the Act describes, and could be claimed by right, while the scheme amount is what the delivery system in fact pays. The distance between the two figures is the distance between the layers, measured in a single benefit.

The largest scheme of the layer is the Ayushman Bharat Pradhan Mantri Jan Arogya Yojana, administered by the National Health Authority, which provides a health cover of five lakh rupees per eligible family per year for hospitalisation, with eligibility drawn from deprivation and occupational criteria in the Socio-Economic Caste Census. It is also the scheme most often mistaken for a right, which is why it earns its own direct answer here.

Is Ayushman Bharat a legal right?

No. Ayushman Bharat Pradhan Mantri Jan Arogya Yojana is an executive scheme run by the National Health Authority, not an entitlement created by an Act of Parliament. Its cover of five lakh rupees per family per year, its criteria and procedures are set by scheme decisions, and a person denied its benefit has a complaint, not a statutory claim.

A scheme has an anatomy, and it is worth learning to read it, because the anatomy is where the programme’s legal weakness is inscribed. A scheme begins as an announcement or a budget proposal, is approved through the expenditure machinery of the government, and is issued as guidelines by the administering ministry: a document that fixes the objective, the beneficiary criteria, the benefit, the funding pattern between the Union and the states, the implementing agency and the monitoring arrangement. It is approved for a period, characteristically aligned to the cycle of a Finance Commission, and its continuation beyond that period requires fresh appraisal and approval, a point at which programmes are routinely restructured, merged or closed. No stage of this life cycle involves a legislature voting on the programme as such; the legislature votes the ministry’s demand for grants in aggregate. Amendment of the programme requires no instrument more solemn than a revised guideline or an office memorandum. Courts asked to enforce a scheme proceed on the footing that its terms bind the administration as a matter of its own declared policy while the scheme subsists, which is a real but derivative protection: it polices departure from the guidelines, and it cannot prevent the guidelines from being changed prospectively or the scheme from being wound up. The beneficiary’s security under a scheme is therefore the security of administrative regularity, valuable and enforceable in its own terms, and categorically less than the security of a text the executive cannot amend.

State governments operate the same layer at their own level, and several of the largest welfare commitments in any state budget are state schemes of identical legal character: income support to farmers layered over the Union’s, pension top-ups above the assistance programme’s rates, free or subsidised services announced in state budgets, and periodic loan waivers effected by executive and budgetary action. The layer’s logic is the same at both levels of the federation, and so is its legal weakness, and a beneficiary comparing a state promise with a Union statute should apply the audit to each in turn rather than assume that the larger government is the one bound by law.

The legal consequences of scheme status are uniform across the layer, and they are the mirror image of the statutory position. Eligibility is whatever the guidelines say it is, and guidelines can be revised between one budget cycle and the next. The beneficiary list is an administrative product, prepared from census or survey data, and an omitted household’s remedy is an application for correction within the programme’s own procedures. No officer owes the applicant a statutory duty with a time limit attached. No allowance accrues when the benefit is delayed or refused. And the programme’s continuation depends on appropriation: a scheme that is not funded in a given year’s budget is not breached, merely discontinued, and no claim survives its discontinuation. None of this makes a scheme unimportant; several schemes reach more people than several statutes. It makes a scheme a different kind of legal object, and the first task in any welfare question is to identify which kind of object one is dealing with.

The federal division: who may legislate, who must deliver

The Constitution distributes the welfare field between the Union and the states in a way that explains the framework’s characteristic shape: Union texts, state delivery. The Concurrent List, on which both Parliament and the state legislatures may legislate, carries the field’s core in two entries: Entry 23, social security and social insurance, and Entry 24, welfare of labour, including conditions of work, provident funds, employers’ liability, workmen’s compensation, invalidity and old age pensions and maternity benefits. The State List carries, in Entry 9, relief of the disabled and unemployable, and the states hold the subjects through which welfare is physically delivered: public health, local government, land and water. The Union List gives Parliament the inter-state and residuary routes and, decisively, the greater revenue heads. The structural result is the one visible across the whole map. The Union legislates the framework statutes for the largest entitlements and finances the larger share of them, because it commands the elastic taxes. The states identify the beneficiaries, staff the machinery, run the institutions and answer for delivery, because the Constitution places the delivery subjects and the local state in their hands. A welfare statute in India is therefore characteristically a Union promise administered by a state, and many of its failures are failures of the joint between the two: the Union’s text assumes an identification the state performs badly, the state’s machinery awaits a release the Union delays, and the beneficiary’s remedy has to identify which half of the joint failed before it can be aimed.

How the money flows: four funding architectures

The funding question in welfare law is usually asked too loosely, as if the only issue were the size of the allocation. The legal structure of the funding matters as much as its size, because the structure decides whose obligation the money is. The framework displays four distinct architectures. The first is the statutory funding share, in which the Act itself divides the cost between the Union and the states: the employment guarantee’s division, under which the Union bears the full unskilled wage cost and three-fourths of the material cost, and the food statute’s arrangement, under which the Union is responsible for allocating the foodgrains the entitlement requires and shares the costs of the distribution system. A funding share written into a statute cannot be altered by a budget circular; it is part of the entitlement’s legal design, and a state that is not reimbursed in accordance with it holds a statutory grievance of its own.

The second architecture is the budgetary scheme, in which the whole commitment rests on annual appropriations voted by the legislature as part of the general budget. Within this architecture there is a further division that beneficiaries rarely see: between centrally sponsored schemes, whose cost is shared between the Union and the states in a ratio fixed by executive decision, conventionally sixty to forty for most states in the period after the Fourteenth Finance Commission raised the states’ share of central taxes, and central sector schemes borne wholly by the Union. The ratios are policy, not law. They have been revised by executive action and can be revised again, and a state that declines or delays its share can stall a centrally sponsored programme inside its territory without breaching any statute. The third architecture is the fund: the insurance corporations of the 1948 and 1952 Acts, the cess funds of the welfare boards, and the social security fund contemplated by the Code on Social Security, each a pool fed by identified contributions and dedicated to identified purposes. Fund money is harder to divert than budget money and easier to leave unspent, since a fund’s benefits move only when its administrators frame and operate the measures the fund exists to finance. The fourth architecture is no public money at all: the senior citizens statute’s maintenance, which is a private obligation between family members that the state enforces through a tribunal, and the reservation provisions of the disability statute, which distribute opportunities rather than expenditure.

A constitutional provision underlies the budgetary architecture and explains its flexibility: Article 282 empowers the Union and the states to make grants for any public purpose, notwithstanding that the purpose is outside their legislative competence. The great scheme transfers of the decades of central planning and the centrally sponsored schemes of the present framework rest substantially on this spending power rather than on any legislative entry, which is why a programme can be national in coverage without being statutory in form, and why its conditions can be prescribed by the granting government as terms of the grant. The food subsidy displays the budgetary architecture at its largest: the difference between the economic cost of procuring and distributing grain and the issue prices the food statute fixes is borne by the Union as food subsidy, so that the Act’s three, two and one rupee prices are maintained by an annual appropriation standing behind the statutory text, a reminder that even the most completely statutory entitlement in the field is, at the point of payment, a charge on a budget that a legislature must vote every year.

Who pays, the question the aspirant is asked and the beneficiary needs answered, therefore has four different kinds of answer in this field, and each answer predicts a different failure. Statutory shares fail by intergovernmental dispute and delayed release. Budgetary schemes fail by appropriation, silently, in the estimates of a single year. Funds fail by accumulation without delivery, when contribution outruns registration and administration. Private obligations fail by the poverty of the family on which they are laid, which no tribunal order can cure. A welfare analysis that reports only the total expenditure on a programme, without identifying its architecture, has not yet begun to describe the programme’s legal reality.

Identification: the real gate of the welfare state

Every instrument in the framework must answer a question prior to all others: who, exactly, is the beneficiary? The methods by which Indian welfare law answers it differ so sharply that the method, as much as the benefit, determines the lived character of the programme. The employment guarantee uses self-selection: the benefit is a day of manual work at the statutory wage, any rural household may register and demand it, and the design assumption is that the nature of the benefit screens the applicants. The food statute uses census-based identification: states identify priority and Antyodaya households within the coverage percentages the Act fixes, working from the Socio-Economic and Caste Census and their own criteria, and the Act designates the eldest woman of eighteen years or above in the household as its head for the issue of the ration card. The housing and health-cover schemes use the same census base and its deprivation criteria. The disability statute uses certification: the benefit follows a medical authority’s certificate, and the benchmark category follows the forty per cent threshold, so the gate stands in a hospital rather than in a ministry. The forest statute uses deliberative identification, a claim initiated and verified in the Gram Sabha. The street vending statute uses survey identification by a committee on which the surveyed population itself sits. The unorganised sector codes and schemes use registration, of which the national database of unorganised workers created by executive action is the largest instrument: a registry entry becomes the practical precondition of benefits that the statute or scheme then defines.

Identification is where the two layers of the framework most often exchange places in practice. A statutory entitlement with a defective identification exercise delivers less security than its text promises, since the unidentified household holds a right it cannot present; the food statute’s coverage percentages make the point in law, because the Act guarantees the share and the state decides the names. A scheme with a clean, portable registry can deliver with a regularity that compares well with a badly identified entitlement. Reform writing about this field concentrates on rates and coverage, but the statutes themselves show the deeper pattern: Parliament has legislated most carefully where it also legislated the identification method, and the grievance machinery of the rights statutes is occupied to a great extent with identification disputes, the name wrongly excluded, the certificate wrongly refused, the household wrongly classified. The beneficiary definition in the text and the beneficiary list in the office are two different documents, and welfare law lives in the distance between them. At the base of the identification structure sits the certificate, the smallest instrument of the welfare state and the one a citizen meets first: the income certificate, the caste certificate, the disability certificate, the residence proof, each issued by a local authority under state practice and each capable of deciding a claim before any statute is opened. A certificate wrongly refused or wrongly delayed is, for most welfare purposes, indistinguishable from the benefit itself being refused, and the remedies against the certificate’s refusal are thinner than the remedies the parent statutes provide for the benefit, because the certificate is usually an executive product even when the benefit it gates is statutory. Reform of identification in this field is therefore mostly reform of the certificate layer, and the movement toward self-declaration, digitised records and deemed approval in service delivery law is an attempt to push the gate back toward the statute’s own definition, where it can be seen and challenged.

Delivery technology has narrowed some of that distance without changing any legal position. The transfer of cash benefits directly into beneficiaries’ bank accounts, generalised as administrative policy from 2013 onward, and the seeding of beneficiary databases against identity records, altered the mechanics of payment across both layers alike: leakage points were removed, duplicate and ghost entries were struck off, and a payment failure became traceable to a named account and a named transaction. None of that converted a scheme into an entitlement or an entitlement into a scheme. The legal source of a benefit is unaffected by the pipe through which it is delivered, and a beneficiary whose payment stops has exactly the remedy, claim or complaint, that the source of the programme gave them before the pipe was built. The distinction this article draws is therefore not dissolved by administrative modernisation; it is, if anything, sharpened by it, because efficient delivery makes the remaining difference, the difference of legal status, stand out as the whole of the matter.

Reading a welfare statute: the anatomy to look for

The audit of the last section works because welfare statutes share an anatomy, and the reader who knows where each element sits in a text can classify a statute from a single reading. The definitions come first and do more work than any other part: the definition of the beneficiary, household, family, worker, disability or vendor, fixes the population of the right, and the exclusions written into a definition are as operative as the inclusions. The entitlement provision follows, and its verb is the tell: a provision that a person shall be entitled, or that the state shall provide, creates the claim, while a provision that the government may frame schemes for a class creates a power, and the distance between those verbs is the distance between the layers. The machinery provisions name the officers and bodies, Programme Officer, District Grievance Redressal Officer, Maintenance Tribunal, Commissioner, and the presence of a forum with a named presiding officer is the mark of a legislature that intended the entitlement to be used. The funding provisions state the share, the fund or the appropriation route. The audit provisions, social audit, vigilance committee, public record, disclose whether the legislature expected the delivery to be watched. And the closing provisions, on rule-making, removal of difficulties and repeal, show what the statute displaced and how much of its own content it delegated to the executive that administers it, for every welfare statute delegates something, and the proportion delegated is the measure of how much of the scheme layer survives inside the statutory form. The food statute delegates identification. The disability statute delegates its social security content through a qualifying clause. The social security code delegates nearly everything except the definitions. Reading the anatomy in this order converts the classification this article teaches from an impression into a procedure. Applied to the senior citizens statute, the procedure runs in seconds: the definitions fix the protected class by age and relationship; the operative provisions use the language of obligation but lay it on children and relatives; the machinery is a tribunal of the summary kind; the funding provision is absent because the money is private; the audit provisions are vestigial; and the conclusion follows, that the Act is a complete entitlement apparatus pointed at the family, leaving the state’s own provision, the old age home and the medical support, in the softer language of direction. A statute that looks at first glance like a welfare payment law is revealed by its anatomy as a maintenance jurisdiction, and the reader who has performed the exercise once will not again mistake the presence of an Act for the presence of a state promise to pay.

At the district, all of the anatomies converge on one office. The district administration, under the Collector, is the nodal point at which the employment programme’s coordinator sits, the food statute’s grievance officer is appointed, the disability certificate’s issuing authority is supervised, the assistance programme’s lists are approved and the welfare board’s district machinery reports. The convergence is administrative rather than legal, no statute creates it, but it explains a feature of welfare experience that the texts alone do not: a citizen’s welfare business of every legal kind is transacted in the same compound, often in the same queue, and the distinctions this article has drawn, invisible in the architecture of the office, have to be supplied by knowledge. That is the final justification for teaching the classification at all. The state presents its welfare face as one counter. The law behind the counter is several different laws, and the citizen who knows which one governs the matter in hand is the only one the counter cannot turn away with an answer meant for a different programme.

Act or order: the field in one table

The table below is the working instrument of this article. It places the major programmes of the framework side by side and records, for each, the five facts that decide a beneficiary’s legal position: whether the programme rests on a statute or on an executive decision, who the defined beneficiary is, whether a statutory grievance officer or ombudsman exists, whether the Union is obliged to fund it, and what a person can do when the benefit is denied. The final column is the point of the exercise, and it divides the table into two halves that share nothing except the word welfare. Readers who keep the table for reference can open it as a document through the ReportMedic office file viewer alongside their own notes on the statutes it lists.

Programme	Act or executive decision	Defined beneficiary	Statutory grievance officer or ombudsman	Union funding obligation	When it is denied
Rural employment guarantee	Act: Act 42 of 2005	Adult members of rural households volunteering for unskilled manual work	Programme Officer and District Programme Coordinator under the Act; district Ombudsman under Union instructions	Union bears the full unskilled wage cost and three-fourths of material cost under Section 22	Application carries a dated receipt; unemployment allowance accrues under Section 7; an enforceable claim
Food security entitlement	Act: Act 20 of 2013	Priority and Antyodaya households identified by states; pregnant women, lactating mothers and children for specified benefits	District Grievance Redressal Officer under Section 15; State Food Commission under Section 16	Union allocates foodgrains to the states under the Act and shares the cost of the system	Food security allowance under Section 8 where supplies fail; complaint to the District Grievance Redressal Officer
Free and compulsory education	Act: Act 35 of 2009	Every child between the ages of six and fourteen	Local authority under the Act; monitoring by the commissions for protection of child rights	Cost shared between the Union and the states under the Act	Complaint to the local authority and the State Commission; an enforceable duty
Forest rights recognition	Act: Act 2 of 2007	Forest dwelling Scheduled Tribes and other traditional forest dwellers	Gram Sabha and the Sub-Divisional and District Level Committees; petition within the committee structure	None of the spending kind; the Act recognises rights rather than funding a benefit	Petition within the committee structure; a claim to recognition, not to payment
Disability rights and reservation	Act: Act 49 of 2016	Persons with disabilities; benchmark disability of forty per cent for the reserved benefits	Grievance Redressal Officer in government establishments under Section 23; Chief and State Commissioners	No general funding entitlement; duties of access, reservation and non-discrimination	Complaint to the establishment’s officer and to the State Commissioner; enforceable
Senior citizens maintenance	Act: Act 56 of 2007	Parents and senior citizens unable to maintain themselves	Maintenance Tribunal and Appellate Tribunal under the Act	Maintenance is payable by children or relatives, not from Union funds; states carry the institutional duties	Application to the Maintenance Tribunal; an adjudicable claim against the family
Transgender persons protection	Act: Act 40 of 2019	Transgender persons, holding a certificate of identity from the District Magistrate	National Council for Transgender Persons, with a grievance function; no individual benefit officer	Governments are obliged to formulate welfare schemes; no funding share for a defined benefit is fixed	Complaint under the Act for discrimination; welfare benefits depend on the schemes framed under it
Street vending protection	Act: Act 7 of 2014	Street vendors identified in the Town Vending Committee survey and holding certificates of vending	Committee under Section 20 chaired by a person who has been a civil judge or a judicial magistrate; appeal to the local authority	Regulatory statute; no Union funding obligation for a benefit	Application to the Section 20 committee; enforceable protection within the Act’s terms
Social security for unorganised and gig work	Act: Act 36 of 2020, operating through schemes framed under it	Unorganised, gig and platform workers as defined in the Code, on registration	The machinery of the scheme framed under the Code and the boards the Code continues	A social security fund under the Code, fed by contributions that may include aggregator contributions	Determined by the scheme framed; the Code empowers and the scheme defines
National Social Assistance Programme pensions	Executive programme, 1995	Below poverty line elderly persons, widows, persons with disabilities and bereaved families, per the programme guidelines	None created by statute; programme administration only	Central assistance at rates fixed by the guidelines; states add shares of their own	Application and administrative complaint; no statutory claim
Ayushman Bharat health cover	Executive scheme, National Health Authority	Families meeting the scheme’s deprivation and occupational criteria	Scheme grievance channels; no statutory officer	A budgetary commitment under the scheme	The scheme’s grievance process; a complaint, not a claim
PM-KISAN income support	Executive scheme, 2019	Landholding farmer families per the scheme’s guidelines and exclusions	Programme administration, with correction through its own procedures	A budgetary commitment under the scheme	Correction and complaint within the programme; no statutory claim
Housing assistance	Executive scheme	Households meeting the housing and deprivation criteria of the guidelines	Programme administration only	Budgetary, shared with states under the programme’s design	Application and complaint within the programme; no statutory claim
Integrated Child Development Services	Executive programme, 1975	Children under six and pregnant and lactating mothers, served through anganwadi centres	Programme administration; the nutritional entitlements of the covered categories carry the food security statute’s machinery	Budgetary, shared with states	A divided position: the services rest on the programme, the nutritional floor for covered categories rests on the food security statute

Two cells of the table reward a second look because they mark the boundary rather than a side of it. The employment guarantee’s ombudsman is a statutory programme’s officer who rests on Union instructions rather than on the Act itself, a small executive graft on a statutory body, inserted because the Act’s own officers are the programme’s administrators and an appeal to them against themselves was thought insufficient. The child development row is divided against itself, the services resting on a programme of 1975 and the nutritional floor beneath two of its beneficiary classes resting on the food statute of 2013, so that a single anganwadi worker delivers, in one morning’s work, benefits of two different legal kinds to children who cannot tell them apart and whose guardians may one day need to. Boundaries in this field do not run between programmes; they run through them.

Read down the final column and the architecture of the field becomes visible. Every row in the statutory half ends in a forum: a tribunal, a committee, a grievance officer, a commissioner, an allowance that accrues by operation of the Act. Every row in the scheme half ends in a procedure internal to the programme itself. The difference is not the generosity of the benefit. Some scheme benefits are larger than some statutory ones, and some statutory duties, as the senior citizens row shows, are owed by a family rather than by the treasury. The difference is the identity of the decision-maker who can change the position: a legislature that must pass an amending Act in public, or an executive that may revise a guideline by notification.

The politics that made the statutes

Statutes of this kind are not produced by drafting technique, and the rights turn had identifiable political mechanics that a map of the field should record, because they predict where the next statute is likely to appear. Each enactment of the turn stood behind a campaign that had already done the work of definition: the employment guarantee behind a movement for the right to work that could point to a working state model, the food statute behind a right to food campaign that had litigated the programme layer for a decade and arrived with the entitlement’s content substantially argued out, the forest statute behind a mobilisation of forest dwellers against eviction, the education statute behind a longer campaign that had already won the constitutional amendment. The National Advisory Council of the period supplied a drafting channel through which several of these texts passed into government. The pattern matters for prediction. Contingencies acquire statutes when three conditions coincide: a constituency organised enough to sustain a claim over years, a text ready enough to be enacted when the political window opens, and a demonstration, from a state, a court or a programme, that the entitlement can be administered. Contingencies lacking any one of the three remain in the scheme layer regardless of the strength of the moral claim, which is the political form of the legal observation that the directive principles bind no one until a legislature is moved to embody them.

The persistence of the scheme layer is not a failure of understanding on the part of governments that use it, and the stance of this series requires the executive’s case to be heard in its own terms. A scheme can be announced in a budget speech and be paying beneficiaries within the year, where 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 lets a government respond to a newly measured need, a drought, a price shock, a pandemic, at the speed of the event. A scheme’s criteria can be tuned to a state’s conditions by the government that knows them, and its scale can be matched to the year’s revenues without the embarrassment of amending a statute downward. Ministers defending the scheme method make each of these points, and each is true. The beneficiary’s insecurity is not a refutation of them; it is the price of them, and an honest account of the field records both the price and what it buys, leaving the judgment between speed and security where the Constitution left it, with the political process, informed by the knowledge of what each instrument is.

Two models, two failure modes

What did the rights turn of the 2000s change?

It changed the legal character of the largest welfare programmes. Benefits that had been distributed as administrative programmes, among them rural employment, subsidised food, elementary education and forest tenure, were re-enacted as statutory entitlements with beneficiaries defined in the Act, duties placed on named authorities, grievance machinery attached, and funding shares written into the statute itself.

Two arguments dominate discussion of this architecture, and each is usually made in a form too strong for the record. The first holds that the statutory entitlement is simply the superior instrument, because it converts a supplicant into a rights-holder. The record qualifies that claim in a specific way. Several of the entitlements are formally open-ended and operationally rationed. The employment guarantee promises one hundred days to every rural household that demands work, yet the programme operates within an annual approved labour budget, and the food security entitlement, though defined by population share, depends on identification exercises and on the fiscal settlement between the Union and the states. The entitlement fixes the legal position of the person inside the system; the appropriation decides how much of the promise the system can honour in a given year, and the employment and food articles in this series document that gap from within their own statutes. A right that is rationed is not worthless, since the rationing can itself be challenged and the individual refusal remains reviewable, but the claim that enactment alone secures delivery is one the record does not support.

The second argument holds that schemes are the better instrument because they are flexible: a government can design a programme for a newly visible need, target it precisely, revise its criteria as evidence arrives, and spend at the scale the budget allows. That claim is also true and also incomplete, because the same flexibility that permits revision permits withdrawal. A scheme can be narrowed, merged, renamed or closed without any legislative act, and its beneficiaries acquire no security against any of those events; the history of Indian rural employment before 2005 is a sequence of programmes, each announced as a commitment and each replaced by a successor with a new name, under which no worker could claim a single day of work by right. Flexibility and insecurity are the same property described from opposite ends.

The honest position is that the two models fail in different ways and that neither is a substitute for the two things both require: adequate funding and functioning delivery machinery. The statutory model fails by rationing and by the gap between the defined beneficiary and the identified one. The scheme model fails by discontinuity and by the absence of any forum outside itself. The systematic comparison of the two models, programme by programme and failure by failure, belongs to the article on rights-based and scheme-based welfare, which owns that question; this article’s conclusion is the prior one, that a reader who cannot tell which model a programme belongs to cannot assess either its promise or its risks.

The courts after the rights turn: supervision of statutes

Once the entitlements existed in statutory form, the character of welfare litigation changed, and the change confirms the architecture this article describes. The PUCL proceedings had required the Court to construct obligations out of the government’s own schemes and the right to life, because no food statute existed. After 2013, litigants could plead the statute. In Swaraj Abhiyan v Union of India, decided in 2016 in proceedings arising out of drought in several states, 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 addressed the employment guarantee’s operation in the drought districts, treating the Acts as the measure of the governments’ obligations. The contrast with the earlier phase is exact. Construction gave way to enforcement: the questions became whether the statutory officer had been appointed, the statutory commission constituted, the statutory allowance paid, and those are questions a court can answer from the text without designing a programme.

The same pattern appears in the litigation over delayed wages and compensation under the employment statute, where the entitlement and the allowance are pleaded from the sections themselves, and in challenges to exclusion from beneficiary lists, where the claimant’s case is that a statutory identification duty was not honestly performed. None of this litigation is costless or fast, and the writ jurisdiction of the High Courts under Article 226 remains a remedy of last resort for a person whose need is immediate. But the pleading has changed its foundation, and with it the distribution of argumentative burden: the beneficiary under a statute does not have to persuade the court that the benefit ought to exist, only that it exists and was denied. That inversion of burden is the single most concrete difference the source of a benefit makes, and it is invisible in any account of the field that counts programmes without classifying them.

Accountability for the framework is not exclusively judicial, and the audit institutions complete the picture the courts begin. The Comptroller and Auditor General has subjected the major programmes of both layers to performance audit, the employment guarantee and the assistance programme among them, and the reports perform for the legislature the function the grievance machinery performs for the beneficiary: they measure the delivery system against the standard its own instrument sets. The difference between the layers appears here too. An audit of a statutory programme can find a breach, because the Act supplies the standard; an audit of a scheme can find only a shortfall against the scheme’s own guidelines and targets, standards the audited government itself wrote. Even the Republic’s auditor, in other words, can hold the executive to a welfare statute more firmly than to a welfare scheme, which is the institutional form of the proposition with which this article began.

What the framework leaves out

A map of a field is also a map of its absences, and the welfare framework’s gaps follow its structure closely enough to be predictable from it. Urban poverty has no employment statute: the guarantee is rural by the Act’s terms, and urban livelihoods are served by an executive mission for urban livelihoods and by the street vending statute’s protection of an existing occupation, which secures a vendor against eviction but creates no claim to work. Domestic workers, among the largest occupations in the country, have no central statute of their own and appear in the framework at the edges of the unorganised workers legislation and state welfare boards. Agricultural labour is reached by the employment guarantee as an employer of last resort rather than as an occupation with its own protections. Homelessness and destitution without a family, a village record or a certificate remain the territory of schemes, night shelters run by urban missions, and discretionary relief, which is to say the territory where the framework began. The migrant worker exposes a further gap that is geographic rather than categorical: most instruments of the framework are administered by the state of residence and were designed for a settled population, so that the worker who crosses a state line may fall out of the beneficiary lists of both states. The response in the food system, the portability of the ration card across the country, was achieved as an administrative and technological arrangement within the statute’s framework rather than by amendment, and it stands as a useful reminder that some defects of the framework are defects of its registers rather than of its rights, curable by administration where the will exists, while others, the absent statute and the absent forum, are curable only by legislation. Inside the covered contingencies, the categories themselves draw lines that the framework rarely defends in terms. The assistance programme’s widow pension is confined by the guidelines to a defined age band, so that a widow below the band holds no pension claim of any kind; the disability pension turns on a severity threshold in the guidelines that differs from the disability statute’s benchmark; the maternity scheme pays for the first living child where the food statute’s benefit is not so confined. Each line is an administrative choice, defensible or not on fiscal grounds, but none of them was debated as legislation, and a beneficiary excluded by one of them has no text to argue from except the guideline that excluded her. The pattern in the gaps is the pattern in the whole: the contingencies that acquired statutes are those around which an organised constituency, a judicial proceeding or a political campaign formed, and the contingencies that remain in the scheme layer are those whose claimants are hardest to organise and easiest to overlook. The law of welfare in India is therefore also a record of which kinds of need succeeded in becoming visible to the legislative process.

The reform debate: consolidation, cash and universality

Three reforms are proposed for this field with sufficient regularity that a map must record them, in the terms their proponents use. The first is consolidation: the argument that the accumulation of statutes, codes, boards and schemes should be rationalised into a coherent social security architecture with a single registry and portable benefits. The labour codes are the partial enactment of this argument in the employment field, and its critics answer that consolidation by codification has repeatedly preserved the scheme method inside the new text, so that the architecture is renamed faster than it is unified. The second is substitution of cash for kind: the food statute itself contemplates cash transfers in place of grain as a reform of delivery, pilot schemes have tested the substitution, and the direct transfer of scheme payments into bank accounts has habituated administrators to cash as the default medium. The proponents of substitution argue from choice and leakage; its critics argue that cash exposes the beneficiary to price fluctuation and to the absence of the physical supply system the entitlement was built on, and that a ration is a guarantee of food in a way a transfer of its cash equivalent is not. The third is universality: the Economic Survey of 2016 and 2017 devoted a chapter to examining a universal basic income, presenting it as a conceptually appealing response to the misallocation and exclusion errors of targeted programmes, while recording within the same discussion the fiscal scale of any meaningful universal payment and the risk that a universal floor would be financed by withdrawing the targeted provisions that serve the poorest. None of the three reforms has been enacted as a general settlement. Each is best understood as a position on the question this article has kept at the centre: not how much the state should spend, but in what legal form its promises should be made, and to whom the promise should be enforceable.

The delivery tier: where the framework meets the citizen

Between the statute book and the beneficiary stands a delivery tier that the framework staffs from the bottom of the administrative pyramid, and the legal character of a benefit is often decided there rather than in the text. The employment guarantee is executed substantially through the Gram Panchayat, which the Act charges with executing at least half the works by cost and whose Gram Sabha conducts the social audit and recommends the works, so that the lowest democratic institution is simultaneously the programme’s contractor and its auditor. The food system terminates in the fair price shop, run not by an official but by a licensee whose licence the state grants and may cancel, and whose obligations to the cardholder are statutory in origin and commercial in daily experience. The anganwadi worker who delivers the nutrition programmes is an honorary worker under a scheme, a status the courts have been asked to convert into employment without success, so that the person who embodies the welfare state for the young child is herself outside the employment protections the welfare statutes give others. The accredited social health activist of the health missions stands in the same position. A map of welfare law that stops at the statute and the ministry leaves out the tier at which entitlements are actually honoured or lost, and at which the grievance machinery described next must operate if it is to operate at all.

The grievance machinery almost nobody uses

A feature of the statutory half of the table deserves separate attention because it is the least visible part of the design. Parliament, when it enacts a welfare entitlement, characteristically builds a forum into the statute: a District Grievance Redressal Officer and a State Food Commission for food, Maintenance Tribunals for senior citizens, a judicially chaired committee for street vendors, grievance officers and commissioners for disability, social audits conducted by the Gram Sabha for employment works, and committees of the forest dwellers’ own institutions for forest rights. These bodies share a design intention: to give the beneficiary a remedy cheaper and faster than a writ petition, in a forum that sits near the delivery point and can order the local authority to act.

The intention has been only partly realised, and the reasons are structural rather than accidental. The existence of the forum is stated in the statute but is rarely stated at the point of delivery, so the person refused a ration, a certificate or a maintenance order frequently does not know the forum exists. Several of the bodies depend for their constitution on state action that the Act invites rather than compels on a timetable, so the forum may exist on paper in one state and in fact in another. And the officers who staff the machinery are usually drawn from the same administration whose decision is being challenged, which limits the distance a complainant can expect. One institution stands across both layers and partly compensates for the machinery’s invisibility. The Legal Services Authorities Act, 1987 (Act 39 of 1987) entitles specified classes, including members of the Scheduled Castes and Scheduled Tribes, women and children, industrial workmen and persons with disability, to free legal services, and maintains authorities at the national, state and district levels whose functions include legal awareness and assistance in accessing government benefits. The legal aid clinic and the district authority’s front office are, in practice, among the few places where a refused beneficiary can learn which of the two layers their benefit belongs to and what forum the source provides, which is the first piece of knowledge every remedy in this field requires.

The result is a pattern visible across the field: the statutory remedy is real, formally complete, and lightly used, while the heavily used remedy remains the political one, an application pressed through a representative or an official’s camp office, which is precisely the channel the statutory design was meant to supersede. For a person deciding what to do about a refusal, the general account in this article is information about the legal structure and not advice on an individual case, where the forum, the limitation position and the facts of the refusal decide the route.

How to identify the source of any programme

Who runs the welfare statutes: the Centre or the states?

Both, in a division the statutes themselves draw. Parliament enacts the framework statutes and the Union funds the largest shares of the spending entitlements, while identification of beneficiaries, delivery and the constitution of much of the grievance machinery fall to the states, so a welfare statute is usually a Union text operated by a state administration.

The practical skill this article exists to teach is a short audit that can be run on any programme a reader encounters, in a speech, a budget speech summary, a poster at a government office or a news report. The audit asks five questions in order. First, is there an Act? A programme with a statute behind it will be traceable to a named Act of Parliament or of a state legislature, with a year; a programme described only by a mission name, a yojana title or a departmental acronym almost always rests on executive decision. Second, does the instrument define its beneficiary, in a section or only in a guideline? A statutory definition can be changed only by amendment; a guideline definition can be changed by the ministry. Third, does it name a duty-bearer and a time limit? Entitlements characteristically attach a duty to a specified officer or authority and a period within which the duty must be performed, as the employment statute does with its fifteen days. Fourth, does it create a forum? The presence of a grievance officer, tribunal, commission or committee constituted by the instrument itself is the strongest single marker of the statutory model. Fifth, where does the money come from: a funding pattern written into the Act, or an annual budget line that the programme competes for like any other head of expenditure?

The audit takes less time than it sounds. Run it on the income support payment to farmer families: there is no Act, the beneficiary is defined in guidelines and revised through them, no officer owes a duty within a stated period, no forum is created, and the money is an annual budget line; the programme sits wholly in the scheme layer, and every one of its features, its rapid launch, its wide coverage and its exposure to revision, follows from that placement. Run it on the employment guarantee: an Act of 2005, a beneficiary defined in the text, a duty with a fifteen-day period, forums and an audit built into the statute, and a funding share in Section 22; the programme sits wholly in the statutory layer, and its features, its litigability, its administrative weight and its resistance to quiet retrenchment, follow from that placement with the same necessity. Between the two poles lie the hybrids the audit exists to catch: the statute that delivers through schemes, the scheme that a court’s orders have hardened into an entitlement in practice, the programme whose nutrition component a later Act has placed on a statutory floor while the rest of it remains guideline law.

The answers place the programme on the map, and the map then answers the questions that follow: whether a refusal can be appealed outside the programme, whether the benefit survives a change of government or of guidelines, whether its criteria bind the officer who applies them, and whether a court can be moved on the strength of the instrument alone. Programmes also move between the layers, and the audit detects the movement: the midday meal crossed from scheme to statutory floor through litigation and then legislation, the social security code converted a set of insurance statutes and an enabling Act into a code that still delivers through schemes, and proposals recur to convert pension provision from guideline to statute, which would move the largest scheme population in the country into the statutory half of the table. The framework is therefore best understood not as two fixed lists but as two legal conditions between which programmes migrate, usually in one direction, from discretion toward entitlement, and usually after a campaign, a judgment or a fiscal crisis has made the discretionary position untenable.

The state layer and the behaviour of the two models under pressure

The framework described so far is the Union’s, but welfare is a subject on which the states legislate and spend on a comparable scale, and the map is incomplete without them. Social security and welfare of labour fall in the Concurrent List, and relief of the disabled and unemployable appears in the State List, so state legislatures hold an independent competence they have used in both models: state welfare fund Acts of the kind Kerala pioneered for occupational groups, state pension and assistance schemes that in several states pay more than the Union’s guideline rates and reach beyond the below poverty line category, and, extending the statutory model into the newest territory of the field, the Rajasthan Platform Based Gig Workers (Registration and Welfare) Act of 2023, the first state statute to create a welfare board and a cess on aggregators for platform workers, with Karnataka legislating on the same subject after it. A beneficiary’s full position is therefore often a composite: 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 audit described in this article has to be run on each layer separately.

Fiscal pressure is the test that separates the models in operation, and the brief for this field rightly insists on it. When revenues contract or priorities shift, the two layers absorb the shock differently. A scheme absorbs it immediately and invisibly: the allocation is trimmed, the instalment delayed, the eligibility circular revised, and no public act records the change. A statutory entitlement absorbs it slowly and contentiously: the legal duty survives the budget, the funding share remains in the Act, arrears and unpaid allowances become litigable, and the reduction, if it is to be made lawfully, must be made by amendment in the open or endured as a gap between duty and delivery that courts and auditors can name. This is the practical content of the proposition that a right is more secure than a scheme. The security is not that the money cannot fall short; the rationing documented in the employment and food statutes proves that it can. The security is that the shortfall is visible, attributable and challengeable, because the standard against which it falls short is written in a public text that the executive did not write and cannot rewrite.

The verdict on the field

India’s welfare law is not a system, and describing it as one is the error from which most confused accounts begin. It is an accumulation: a constitutional instruction that no court can enforce, a judicial bridge that carried parts of the instruction into the enforceable chapter, an insurance model built for a workforce that most workers never joined, a family-duty model for old age, a decade of rights statutes that converted the largest programmes into entitlements, a consolidating code that kept the scheme method inside statutory dress, and a scheme layer that continues to grow alongside all of it because it is faster to launch than a bill is to pass. Each layer answers the question its predecessor left open and leaves a question of its own. What unifies the field for the reader is not a common design but a common test. For every benefit the state offers, there is a source, and the source decides who is a beneficiary, who bears the duty, who funds it, which forum hears the complaint, and what survives a change of policy. Welfare laws in India become legible, and a denied benefit becomes actionable or recognisably not actionable, at the moment that test is applied.

The three readers this article was built for leave with three different instruments. The person seeking a benefit, and those assisting them, leave with the audit and the table: identify the source, find the forum the source creates, and measure every promise by the remedy attached to it. The researcher leaves with the two-model analysis and the four funding architectures, a structure in which any programme, including one announced tomorrow, can be placed and its behaviour under pressure predicted. The aspirant leaves with the constitutional skeleton and the statute set, which is the part of the field that examination papers test and the part that changes slowest.

For the student of the Constitution, the field repays examination in the form this article has given it, because the examinable structure and the real structure are the same. The directive principles supply the ends, in Articles 38, 39, 41, 42, 43, 46 and 47; the fundamental rights supply the bridge, in the Article 21 cases; the statutes supply the operative law, contingency by contingency, each with its beneficiary, its duty-bearer, its forum and its funding rule; and the schemes supply the standing reminder that most of what the state does in this field it does by choice rather than by obligation. A candidate who can place any named programme on that structure, and say what follows from its placement for the person denied it, commands the subject. A candidate who recites programmes in a list commands only the list, and the list is the one thing about this field that changes every year.

Frequently Asked Questions

Q: Which laws make up India’s welfare framework?

The framework has four layers. The constitutional layer is Part IV of the Constitution, especially Articles 41, 42, 43, 46 and 47, read with the Article 21 judgments on livelihood, food, shelter, health and education. The statutory layer includes the employment guarantee Act of 2005, the forest rights Act of 2006, the education Act of 2009, the food security Act of 2013, the street vendors Act of 2014, the disability Act of 2016, the senior citizens Act of 2007, the transgender persons Act of 2019, the insurance Acts of 1948 and 1952, the construction workers Acts of 1996 and the Code on Social Security of 2020. The third layer is the body of executive schemes, and the fourth is the state layer of state Acts and state schemes.

Q: What is the difference between a right and a scheme?

A right, in this field, is a benefit created by an Act of Parliament or of a state legislature: the statute defines the beneficiary, places a duty on a named authority, frequently creates a grievance forum, and can be changed only by another Act. A scheme is a programme created by executive decision and governed by guidelines issued by the administering ministry. Its criteria, amounts and continuation can be revised by the same executive route that created it, without any legislative proceeding. A person denied a statutory benefit holds a claim that a forum can enforce. A person denied a scheme benefit holds a complaint within the programme’s own procedures, protected by the requirement that the administration follow its guidelines fairly while the scheme subsists.

Q: Which welfare entitlements are legally enforceable?

The enforceable entitlements are the statutory ones. A rural household can enforce the employment guarantee, including the unemployment allowance where work is not provided in time. A household covered by the food security Act can enforce its grain entitlement and claim the food security allowance where supplies fail. A child between six and fourteen can enforce free and compulsory elementary education. Forest dwellers can enforce recognition of forest rights through the Act’s committees. Persons with benchmark disabilities can enforce the reservation and accessibility provisions of the disability Act. A senior citizen can enforce maintenance against children or relatives before a Maintenance Tribunal. Insured workers can enforce their claims on the insurance funds. Scheme benefits, by contrast, are enforceable only to the extent of the scheme’s own guidelines.

Q: What does Article 41 of the Constitution require?

Article 41 directs the state, within the limits of its economic capacity and development, to make effective provision for securing the right to work, the right to education and the right to public assistance in cases of unemployment, old age, sickness and disablement, and in other cases of undeserved want. It is the directive principle at the centre of the welfare framework, and most of the welfare statutes in this article can be read as partial embodiments of it. By reason of Article 37, the principle cannot be enforced by a court against the state directly. Its force is as an instruction to legislatures and governments, fundamental in the governance of the country, which acquires legal teeth when a statute translates it into a defined entitlement or when a court reads its substance into the right to life.

Q: Who pays for a statutory welfare entitlement?

The statute itself answers the question, and the answers differ by design. Under the employment guarantee Act, the Union bears the entire cost of unskilled wages and three-fourths of the material cost, and the state bears the rest and the unemployment allowance. Under the food security Act, the Union is responsible for allocating the foodgrains the entitlement requires and shares the costs of the distribution system with the states. The education Act’s cost is shared between the Union and the states. The insurance entitlements are paid from funds built out of employer and employee contributions. The construction workers’ benefits are paid from a cess on construction. The senior citizens entitlement is paid by children or relatives, not from public funds at all. There is no single payer in this field; there is a funding rule inside each instrument.

Q: How do you enforce a welfare entitlement that is denied?

The first route is the forum the statute itself creates, and it should be used first because it is designed for the purpose: the District Grievance Redressal Officer and the State Food Commission under the food security Act, the Maintenance Tribunal under the senior citizens Act, the grievance redressal officer and the State Commissioner under the disability Act, the committee under Section 20 of the street vendors Act, and the Programme Officer and appeal structure under the employment guarantee. Where the statutory forum fails or does not act, the constitutional remedy remains: a writ petition to the High Court under Article 226 or to the Supreme Court under Article 32, pleading the statute and the denial. The district legal services authority can assist an entitled person in finding and using the forum. A scheme denial has no equivalent ladder; the route is the programme’s own grievance channel.

Q: Which welfare statutes have grievance officers?

The food security Act has the most elaborate structure: a District Grievance Redressal Officer in every district under Section 15 and a State Food Commission under Section 16, with vigilance committees below them. The disability Act requires every government establishment to designate a grievance redressal officer for disability complaints under Section 23, and adds Chief and State Commissioners with inquiry powers. The employment guarantee works through the Programme Officer and the District Programme Coordinator, with a district Ombudsman under Union instructions. The senior citizens Act substitutes tribunals for officers. The street vendors Act creates committees chaired by a person who has been a civil judge or a judicial magistrate. The transgender persons Act assigns a grievance function to the National Council. The forest rights Act routes grievances through its committee structure itself.

Q: Why do welfare laws often miss the unorganised sector?

The oldest welfare statutes were built on the employment relation. Insurance and provident fund coverage attach to establishments above a threshold size and to employees with an employer of record who can be made to contribute and to keep records. The unorganised worker has no such employer: work is casual, seasonal, home-based or self-employed, and the contribution and record machinery has nothing to attach to. Identification compounds the exclusion, because the unorganised worker is the person least likely to hold the certificates and documents the delivery system asks for. Successive instruments have attacked the problem from different sides: welfare boards financed by a cess, an enabling Act for unorganised workers in 2008, and registration-based inclusion under the Code on Social Security, which defines the unorganised, gig and platform worker and makes registration the gateway to schemes framed for them.

Q: Are old age pensions under the National Social Assistance Programme legal rights?

No. The National Social Assistance Programme is an executive programme launched in 1995, not a statute. Its pensions for elderly persons, widows and persons with disabilities in below poverty line households are paid at rates and under criteria fixed by the programme’s guidelines, with the Union contributing a share and the states adding their own. A pensioner under the programme cannot plead an Act in support of the pension, and the rates, the age bands and the categories can be revised by guideline without any legislative proceeding. The pensioner is protected by the doctrine that an administration must apply its own declared policy fairly and without arbitrariness, and by the political weight of a large beneficiary class. Several states pay pensions above the programme’s pattern under their own schemes, which share the same legal character unless a state has legislated its pension, as a small number have done for particular classes.

Q: Can a state legislature create welfare entitlements of its own?

Yes. Social security and social insurance and the welfare of labour are entries in the Concurrent List, on which state legislatures as well as Parliament may legislate, and relief of the disabled and unemployable is an entry in the State List. States have used the competence in both models. Maharashtra enacted a statutory employment guarantee in 1977, decades before the national Act. Kerala built occupation-wise welfare funds by state legislation. Rajasthan enacted a statute for platform based gig workers in 2023, creating a welfare board and a cess on aggregators, and Karnataka legislated on the same subject after it. A state entitlement binds within the state that enacts it, is enforced through the machinery the state Act creates, and stands alongside Union statutes and schemes covering the same contingency, so that a person’s complete welfare position in a state is frequently a composite of instruments from both levels.

Q: How are beneficiaries identified under the food security statute?

The statute fixes the coverage and leaves the identification to the states. The Act entitles up to seventy-five per cent of the rural population and fifty per cent of the urban population, divided between priority households, which receive five kilograms of foodgrains per person per month, and Antyodaya households, which receive thirty-five kilograms per household per month. Within those shares, state governments identify the eligible households, working from the Socio-Economic and Caste Census and from inclusion and exclusion criteria the states frame under the Act. The Act designates the eldest woman of eighteen years or above in the household as the head of the household for the issue of the ration card. Identification lists are state documents, revised by state processes, and exclusion from the list is the commonest ground on which the statutory entitlement fails in practice, which is why the Act’s grievance machinery is largely occupied with identification disputes.

Q: What can a street vendor do if a certificate of vending is refused?

The street vendors Act builds the remedy into the certification system. Certificates of vending are issued by the Town Vending Committee after the survey of vendors the Act requires, and a vendor who has a grievance or a dispute, including over refusal, suspension or cancellation of a certificate, may apply in writing to the committee constituted under Section 20 of the Act, which is chaired by a person who has been a civil judge or a judicial magistrate and includes two other professionals with prescribed experience. The committee inquires into the application and takes steps for redressal. A person aggrieved by the committee’s decision may appeal to the local authority, which must hear the appellant before deciding. The Act separately protects an existing vendor against eviction or relocation until the survey has been completed and certificates issued, so the certification dispute and the eviction threat are legally distinct, and the vendor can resist the second while pursuing the first.

Q: How does contributory social insurance differ from a tax-funded welfare entitlement?

The two instruments create the claim in opposite ways. Under contributory social insurance, of which the employees’ state insurance and provident fund statutes are the Indian exemplars, the worker and the employer pay contributions into a statutory fund during the worker’s working life, and the benefit, medical care, a sickness or maternity payment, a pension, is a claim on that fund arising on a defined contingency. The claim is earned, its amount commonly bears a relation to the contribution record, and it does not depend on the beneficiary being poor. Under a tax-funded entitlement, of which the employment guarantee and the food security entitlement are the exemplars, no contribution is made or required: the beneficiary qualifies by status, by residence, household or condition, and the cost falls on general revenues under a funding rule the Act prescribes. Insurance therefore excludes whoever never entered the contributory relation, while the tax-funded entitlement excludes whoever falls outside its definition of the beneficiary, and much of Indian welfare history is the attempt to cover the population the first instrument cannot reach with instruments of the second kind.

Q: Why does the senior citizens statute place the maintenance duty on family members rather than on the state?

The design follows the older maintenance tradition in Indian law and a judgment about fiscal reach. Personal law had long obliged a person to maintain aged parents, the Hindu Adoptions and Maintenance Act of 1956 states the obligation for Hindus, and the summary maintenance jurisdiction of the criminal procedure codes applied a version of it across communities. The Act of 2007 generalised that obligation, made it enforceable through a dedicated tribunal with a summary procedure, and added protections the older law lacked, including the power to void a property transfer made on a condition of care that was not honoured. The state’s own assumption of the cost was confined to the residual duties: old age homes and medical support for the indigent. Proponents of the design argue that the family’s duty is prior in fact and in the society’s own norms, and that a tribunal is a cheaper and faster instrument than a pension bureaucracy. The design’s limit is equally visible: it delivers nothing to the senior citizen whose family has no means, which is the gap the assistance programme’s pension exists, inadequately, to fill.

Q: How do courts supervise the delivery of welfare entitlements without administering the programmes themselves?

The characteristic technique is continuing supervision of a proceeding that is kept open while the court monitors compliance, of which the food litigation begun in 2001 is the leading example. The court issues interim directions that convert the governing instrument, a scheme in that case and a statute in later cases, into specific obligations: appoint the officer the Act names, constitute the commission the Act requires, serve the meal the programme promises, file a compliance report by a date. In the drought proceedings of 2016 the Supreme Court directed the implementation of the food security statute’s own machinery in the affected states. Commissioners and committees appointed by the court gather the facts the ordinary record does not supply. The court’s discipline is directed at the instrument’s own standards rather than at a programme of the court’s design, which is what distinguishes supervision from administration: the standard is the legislature’s or the government’s text, and the court’s contribution is to make departure from it visible, attributable and correctable.

Q: Why did the Constitution place welfare duties in the directive principles rather than in the fundamental rights?

The choice was deliberate and was defended in the Constituent Assembly. The framers followed the Irish model of directive principles: instructions to the state that would guide legislation and policy without being enforceable in court. Dr B. R. Ambedkar described them as instruments of instruction to whoever came to power, whose sanction was political: a government that disregarded them would answer to the electorate at election time. Two practical reasons supported the design. Welfare provision depends on revenue, and a court cannot appropriate money or judge the pace at which a poor state can build the apparatus of a welfare system. And the principles were meant to bind governments of every complexion to a common direction without freezing the content of policy, which had to remain open to legislative choice. Proposals to make the principles justiciable were considered and not adopted. The consequence is the structure this article maps: the principles state the ends, and enforceable welfare law exists only where legislatures, and occasionally courts working through the fundamental rights, have supplied the means.

Q: Why do welfare statutes define the beneficiary inside the statute itself?

Because the definition is the entitlement’s boundary, and the legislature that creates a right does not leave its boundary to the beneficiary’s opponent. A beneficiary defined in the Act, the rural household, the child between six and fourteen, the person with benchmark disability, can be excluded only by the grounds and processes the Act itself permits, and an officer who excludes on some other ground acts against the text and can be corrected by the statute’s forum or by a court. A beneficiary defined in guidelines can be redefined by the authority that issued them, between one budget and the next, and the excluded person has no external text to argue from. The definitional choice is therefore the single most consequential drafting decision in a welfare statute, and the disputes that reach the forums are disproportionately definitional: whether this household is a household, whether this disability is forty per cent, whether this worker is within the cess’s trade. Parliament legislates the definition because the definition is where the promise is either kept or quietly withdrawn.

Q: What happens to a person’s welfare claim when no statute’s beneficiary definition covers them?

In strict law, there is no residual welfare entitlement to fall back on. The person descends to the layer beneath the statutes: the schemes, whose criteria may cover the case or may not, and beyond the schemes to discretionary relief and charity. Three mitigations exist, and their limits define the gap. The administration must apply any scheme that does cover the case fairly and without arbitrariness, under the equality guarantee. The right to life jurisprudence supplies a floor in the extreme cases, emergency medical care and freedom from starvation among them, which courts have enforced against the state. Legal services institutions can help the person find the instrument, state or Union, statutory or executive, whose definition does reach the case, since the commonest practical failure is not the absence of any provision but the claimant’s ignorance of which of several dozen provisions applies. The gap is structural rather than accidental: a framework built contingency by contingency, statute by statute, will always have interstices, and the persons in them are characteristically those whose contingency has not acquired a constituency, a campaign or a case.

Q: Are the maternity benefit in the food security statute and the maternity benefit scheme payment the same benefit?

No, and the confusion between them is the standard example of the error this article exists to correct. Section 4 of the National Food Security Act, 2013 entitles a pregnant woman and lactating mother to a maternity benefit of not less than six thousand rupees. The entitlement is in the statute: the class is the one the Act describes, the floor is fixed by the text, and the benefit is a statutory claim. The Pradhan Mantri Matru Vandana Yojana is an executive scheme administered by the Ministry of Women and Child Development, which pays five thousand rupees for the first living child, on conditions its guidelines prescribe. The scheme payment is lower than the statutory floor, confined to a narrower class, and alterable by guideline. Treating the two as a single benefit misstates the amount, the class and the remedy at once. The distinction was drawn in the series’ treatment of women’s welfare legislation and is reproduced here because no other pair of instruments in the field sits so close together while differing in every legally relevant respect.

Q: Does a welfare statute bind private actors, or only governments?

Several of the principal welfare statutes bind private actors directly, and the pattern is one of the framework’s defining features. The disability Act imposes non-discrimination and accessibility duties on establishments that include private establishments, and its reservation in education reaches government-aided institutions. The education Act obliges private unaided schools to admit children from weaker sections and disadvantaged groups to at least a quarter of the entry-level class, with reimbursement from the state. The transgender persons Act prohibits discrimination by establishments in employment, education, healthcare and access to services. The employer-liability and insurance statutes place their duties on employers as such. And the senior citizens Act lays its central duty on private persons, the children and relatives of the claimant. A welfare statute is therefore not characteristically a command to a department. It is a distribution of duties across the state, employers, institutions and families, and identifying the duty-bearer is as much a part of reading it as identifying the beneficiary.