Every payslip in the organised sector carries a line that most employees read once and never examine again: a deduction of twelve per cent of wages, matched by the employer, under a statute enacted when the Republic was two years old. The EPF Act 1952 is the legal machinery behind that line. It decides which establishments must run the deduction, how the money is divided between a lump sum savings fund, a monthly pension and a death insurance benefit, what happens when an employer deducts the money and fails to deposit it, and why the balance in the fund stands ahead of almost every other claim when an employer collapses. A member who understands the Act can answer the questions that decide real money: whether the establishment was covered at all, whether the contribution was computed on the right wage base, whether the pension will be reckoned on the wage ceiling or on actual salary, and which of three different funds pays on death in service.

The Act has acquired a second distinction that no other Indian statute shares. The Code on Social Security, 2020 named it for repeal, the Code itself was brought into force, and the repeals of eight companion enactments were notified. The repeal of the provident fund statute was not. India therefore has a retirement savings law that is printed inside a repealing Code as a dead letter and operates in every payroll office in the country as living law, and a reader who consults the Code will reach exactly the wrong conclusion about which of the two documents governs a contribution. The sums at stake are correspondingly large: the fund holds the retirement savings of a substantial part of the country’s organised workforce, its monthly inflow is a visible component of national household savings, and its declared interest rate is a number that finance ministries and wage boards watch because it prices the savings of workers who have no other market instrument. This article sets out the Act provision by provision, traces the contribution from a month’s wages to the three funds, and then states the commencement position with the precision it requires, because that position has been misreported more often than any other fact in Indian labour law.

A statute built in layers: the EPF Act 1952 from ordinance to pension

The Act began as an emergency instrument. An ordinance promulgated on 15 November 1951 first imposed compulsory provident fund contributions on industrial workers, and the Bill that replaced it was passed by Parliament and received the President’s assent on 4 March 1952, entering the statute book as the Employees’ Provident Funds Act, 1952 (Act 19 of 1952). The short title has been amended twice since, once when a family pension was added and once when the deposit-linked insurance arrived, and in its present form it reads as the Employees’ Provident Funds and Miscellaneous Provisions Act, 1952. The number and the year have never changed, and payroll practice cites the Act by them to this day.

The statute of 1952 was narrower than the one in force at present. It applied to factories in the industries named in Schedule I, a list that opened with six industries, and only where fifty or more persons were employed. The threshold figure of fifty was replaced by twenty by the amending Act 46 of 1960, with effect from 31 December 1960, and the reach of the Act widened steadily thereafter as the Central Government notified further industries and classes of establishment under Section 1(3). The architecture of enforcement arrived in stages. The determination and machinery provisions, headed by Section 7A, were inserted by Act 28 of 1963. The Employees’ Family Pension Scheme began in 1971 under a Section 6A inserted for the purpose, adding a survivor benefit to what had been a pure savings statute. The Employees’ Deposit-linked Insurance Scheme followed on 1 August 1976, under a Section 6C inserted by Act 99 of 1976, which also gave the Act its present name. The amending Act 33 of 1988, effective in the main from 1 August 1988, recast the coverage by agreement provision, strengthened the penalties and standardised the contribution rate at eight and one-third per cent. The amending Act 25 of 1996 then substituted a full pension scheme for the family pension scheme with effect from 16 November 1995, converting the second scheme from a widow’s benefit into a retirement pension system. The Act a payroll team applies is therefore a sediment of seven decades: a savings core from 1952, an enforcement spine from 1963 and 1988, a pension layer from 1971 recast in 1995, and an insurance layer from 1976.

When did the Employees’ Provident Fund Scheme itself begin?

The Act was assented to on 4 March 1952, but the first Employees’ Provident Fund Scheme framed under Section 5 was published in the Gazette on 2 September 1952, by notification S.R.O. 1509, and took effect on 1 November 1952. Contributions under the scheme run from that November.

The scheme is the operative document for almost everything a member experiences. The Act creates the obligation, fixes the contribution rate in Section 6 and the scheme-making powers in Sections 5, 6A and 6C, and leaves membership, accounts, withdrawals, nominations and claims to the three schemes framed under it. Section 7 empowers the Central Government to add to, amend or vary a scheme by notification, prospectively or retrospectively, and Section 6D requires every scheme to be laid before Parliament. The result is a two-level statute: a compact Act of twenty-two sections that changes rarely, and three detailed schemes beneath it that change often, each change carrying the authority of the Act behind it. The amending power in Section 7 deserves its own emphasis, because it explains how the schemes could be transformed in 1995 and again in 2026 without a fresh Act for each transformation. A scheme amendment may be retrospective, and the pension substitution of 1995 was made by an amending Act of 1996 operating from November 1995, with the members’ accrued expectations protected not by any bar on retrospectivity but by the express savings in Section 6A(3) for the ceased scheme’s beneficiaries and by the transition rules of the new scheme for serving members. Retrospectivity in the schemes is therefore a governed power: the executive may rewrite the subordinate code backwards, and the courts then measure the rewriting against the parent Act and against the protection the parent Act itself extends to accrued benefits, which is the precise ground on which the 2014 pension amendment was attacked and, in the main, sustained.

Before the Act: the coal mines precedent and the road to 1952

Compulsory provident fund saving did not begin in 1952. Its first statutory form in India was sectoral: the Coal Mines Provident Fund and Bonus Schemes Act, 1948 created a fund for coal miners, administered separately, four years before the general statute. The generalisation followed the logic of the new Constitution’s Directive Principles, which committed the state to securing a right to public assistance in old age and disablement, and it followed a practical sequence as well: a provident fund ordinance was the quickest instrument available to a government that did not wish to wait for the parliamentary session. The ordinance of 15 November 1951 imposed the contribution, and the Act of 1952 repealed and re-enacted it with Parliament’s authority, which is why the scheme framed under the Act could take contributions from 1 November 1952 with the ordinance period counted in the fund’s early history. The six industries of the first Schedule I were the commanding heights of the organised factory sector of the day: cement, cigarettes, electrical, mechanical and general engineering products, iron and steel, paper and textiles made from the named fibres. The statute’s subsequent history is the history of that list lengthening and of the Section 1(3)(b) notification power pulling non-factory employment into the net, until the exception in Indian organised employment was no longer coverage but its absence.

Two features of the 1951 to 1952 design deserve record because they explain the Act’s durability. The first is that Parliament legislated a shell and delegated the substance. The Act fixed the coverage threshold, the contribution power, the enforcement machinery and the fund’s legal protections, and left membership rules, benefit conditions and procedure to schemes that the executive could amend without returning to Parliament. Every major adaptation since, the family pension of 1971, the insurance of 1976, the pension of 1995, the ceiling revisions, the account number, was accomplished at scheme level or by short amending Acts, which is why a statute of twenty-two sections has never needed wholesale replacement and why the Code’s drafters, when they came to replace it, copied its structure rather than redesigning it. The second is that the Act was built tripartite from the start: the Central Board’s composition, with employers and employees represented alongside government, gave the fund’s administration a negotiated character that pure departmental administration would not have had, and the interest rate, the investment pattern and the damages policy have all carried the marks of that negotiation. Voluntary funds in individual mills and companies had existed before 1952, and the better of them were the design’s raw material: the Act did not invent the provident fund, it made the fund compulsory, portable in principle and enforceable, converting a benefit that had depended on the enlightenment of the individual employer into an incident of covered employment itself. That conversion, from concession to entitlement, is the single idea the whole statute serves, and the exemptions for establishments running funds of their own are its negotiated exception rather than a contradiction of it.

Coverage under the EPF Act 1952: the twenty employee threshold and its edges

Section 1(3) is the gate through which every establishment must pass, and its two limbs should be kept distinct. Under clause (a), the Act applies to every establishment which is a factory engaged in an industry specified in Schedule I and in which twenty or more persons are employed. Under clause (b), it applies to any other establishment employing twenty or more persons, or class of such establishments, which the Central Government specifies by notification in the Official Gazette. The Indian labour law framework guide sets out where this threshold sits among the numerical triggers of Indian labour legislation; within this Act, the consequence of crossing it is that coverage attaches by force of the statute, without registration as a condition, and the employer’s duty to contribute runs from the date the establishment first satisfies the test. A proviso to Section 1(3) adds a third route: the Central Government may, after giving not less than two months’ notice of its intention, apply the Act to an establishment employing fewer than twenty persons, specifying the number in the notification.

Two further provisions close the exits that employers have historically looked for. Section 1(5), inserted in 1960, provides that an establishment to which the Act applies shall continue to be governed by it notwithstanding that the number of persons employed falls below twenty at any later time. Coverage is therefore a ratchet: it locks on at twenty and does not release when a workforce shrinks, a rule that prevents the simplest avoidance device of splitting or shedding staff. Section 2A, added in 1960, provides that where an establishment consists of different departments or has branches, whether in the same place or different places, all such departments and branches are treated as parts of the same establishment. The unit whose employees are counted is the establishment as a whole, not the office or the depot that happens to file the return.

Section 1(4) supplies the voluntary route, and it matters because a large part of the covered universe entered by agreement rather than by compulsion. Where it appears to the Central Provident Fund Commissioner, on an application or otherwise, that the employer and the majority of employees in relation to an establishment have agreed that the Act should apply to it, the Commissioner may notify the Act’s application from the date of the agreement or a later date named in it. The present text of the sub-section was substituted by the amending Act 33 of 1988. Voluntary coverage is not a lighter species of coverage: once notified, the establishment is governed by the Act in full, and Section 1(5) holds it there even if the workforce later contracts.

Against this width, Section 16 carves out three classes. The Act does not apply to a cooperative society establishment employing fewer than fifty persons and working without the aid of power. It does not apply to an establishment belonging to or under the control of the Central or a State Government whose employees are entitled to contributory provident fund or old age pension under a government scheme or rule. And it does not apply to an establishment set up under a Central, Provincial or State Act whose employees enjoy corresponding benefits under that Act. Section 16(2) adds a discretionary exemption for a class of establishments where the Central Government considers it necessary or expedient having regard to financial position or other circumstances. The pattern is consistent: the Act yields where an equivalent public arrangement already exists, and holds everywhere else. For four decades Section 16 also carried an infancy protection, under which a newly set up establishment was outside the Act for an initial period of three years; the clause was omitted by the amending Act 10 of 1998 with effect from 22 September 1997, and a new establishment is covered from the day it crosses the threshold. Section 4 completes the coverage machinery by empowering the Central Government to add any industry to Schedule I by notification, which is the route by which the schedule grew from its original six industries to the long list it carries at present.

The human unit of coverage is the “employee” of Section 2(f): any person employed for wages in any kind of work, manual or otherwise, in or in connection with the work of an establishment, who gets wages directly or indirectly from the employer. The definition expressly includes a person employed by or through a contractor, and an apprentice other than one engaged under the Apprentices Act, 1961 or under the standing orders of the establishment. Within the Provident Fund Scheme, the “excluded employee” is the mirror image of coverage: an employee whose wages exceed the statutory ceiling of Rs 15,000 a month at the point of joining is outside compulsory membership, although employer and employee may jointly elect to bring the employee in. The ceiling, which is examined in its own section below, thus operates at the coverage stage as well as at the contribution stage.

Does the Act reach workers supplied by a contractor?

Yes. Section 2(f) includes within “employee” any person employed by or through a contractor in or in connection with the work of an establishment, and Section 6 requires the contribution for such employees as for direct ones. Section 8A lets the principal employer recover the sum from the contractor, by deduction from amounts payable or as a debt.

The counting of the twenty has its own jurisprudence. The question under Section 1(3) is how many persons are employed in the establishment, and the courts count the establishment as an economic unit: persons employed in connection with its work, including those supplied by contractors, count toward the threshold, and units that are in truth one establishment, sharing management, finance and function, are counted together, applying the functional integrality inquiry familiar across Indian labour law. An employer who divides a single operation among letterheads therefore gains nothing if the units work as one. The count is also indifferent to the duration of each engagement in the ordinary case, since the section speaks of persons employed and not of a permanent cadre, which is why casual and seasonal engagement patterns feature in so many coverage disputes, and why the Section 7A officer’s first task in a contested case is to reconstruct the true establishment from wage records, contractor bills and attendance material rather than from the employer’s description of itself.

International workers form a special category within the scheme: employees holding foreign passports, and Indian employees working in countries with which India has a social security agreement, are brought into membership on terms that depart from the ceiling in defined respects, with detachment certificates under bilateral agreements relieving double contribution. The detail belongs to the scheme text; the point of principle is that the Act’s coverage follows the establishment, and the scheme then adjusts the terms by class of employee.

The three schemes of the EPF Act 1952 and the three funds they feed

The Act works through three schemes, and confusion between them is the most common error in payroll writing on the subject. The Employees’ Provident Fund Scheme, 1952 is framed under Section 5. The Employees’ Pension Scheme, 1995 is framed under Section 6A. The Employees’ Deposit-linked Insurance Scheme, 1976 is framed under Section 6C. Each scheme maintains its own fund, each fund is administered by the Central Board of Trustees, and a rupee that enters one fund cannot be spent by another.

The Provident Fund Scheme is the savings engine. Contributions are credited to an individual account in the member’s name, interest is credited annually at the rate the Central Government declares, and the accumulation is payable on the events the scheme specifies: retirement, retrenchment, migration, permanent incapacity, and, on death in service, to the nominee or family. The scheme also permits advances from the account during service for stated purposes, among them medical treatment, marriage and education in the family, and the purchase or construction of a dwelling, so that the fund operates as a constrained savings vehicle rather than a locked box. What defines the scheme is individual ownership of a running balance: the member’s entitlement is the balance in the account, no more and no less, and investment risk and administrative cost are absorbed by the fund’s structure rather than priced against the member.

The Pension Scheme is a different kind of instrument wearing similar clothes. Section 6A, as substituted by the amending Act 25 of 1996 with effect from 16 November 1995, empowers the scheme that provides superannuation pension, retiring pension and permanent total disablement pension to employees, and widow or widower’s pension, children pension and orphan pension to their beneficiaries. It replaced the Employees’ Family Pension Scheme, 1971, which had provided a survivor benefit alone. The substitution protected the transition expressly: Section 6A(3) wound up the family pension scheme, vested its net assets in the new Pension Fund, and guaranteed beneficiaries of the ceased scheme benefits not less than those they were entitled to draw under it. The Pension Fund is fed, under Section 6A(2), by a diversion from the employer’s contribution under Section 6, capped by the section at eight and one-third per cent of the wages, by contributions from exempted establishments, by the inherited family pension assets, and by sums the Central Government specifies after appropriation by Parliament. Pension entitlement is not a balance. It is a formula applied to pensionable service and pensionable salary, so the fund pools longevity risk across members in the manner of a defined benefit arrangement, and the adequacy of the pool is an actuarial question rather than an accounting one.

The Insurance Scheme is the smallest and the most frequently forgotten. Section 6C, inserted by Act 99 of 1976 with effect from 1 August 1976, empowers a scheme providing life insurance benefits to employees of covered establishments. The employer alone contributes, paying into the Deposit-linked Insurance Fund an amount the section caps at one per cent of the wages; the Insurance Scheme fixes the rate at one-half of one per cent. The employee contributes nothing. The benefit, called the assurance benefit, is payable on the death of a member while in service, and is computed from the member’s wages and the balance in the provident fund account, within minimum and maximum limits that the Insurance Scheme prescribes. The design links the payout to the member’s own accumulation, which rewards longer membership, and charges the cost to the employer as a condition of employing covered labour.

The Pension Scheme carries its own discipline of valuation. A defined benefit promise funded by a capped diversion can drift out of balance silently, and the scheme therefore requires the Pension Fund to be valued by an actuary at the intervals the scheme prescribes, with the valuer appointed by the Central Government. The valuation reports are the documents in which the fund’s deficit, the recurring subject of the higher pension controversy, is computed. No corresponding valuation discipline applies to the provident fund, because a defined contribution account cannot be in actuarial deficit: it can only be what its balances are.

Each scheme’s permissible contents are themselves listed in the Act’s schedules: Schedule II for the Provident Fund Scheme, Schedule III for the Pension Scheme and Schedule IV for the Insurance Scheme enumerate the matters a scheme may govern, from the persons and wages covered to the conditions of withdrawal, the manner of payment and the penalties for defaults under the scheme. The schedules are the boundary of the executive’s scheme-making power, and a scheme provision outside them would be outside the Act, which is the ultra vires ground on which the 1.16 per cent condition of the 2014 pension amendment eventually fell.

The distinction between the three funds explains most disputes that reach the organisation. A claim for the lump sum is a claim against an account. A claim for pension is a claim against a formula and a service record. A claim on death in service may engage all three at once: the provident fund balance to the nominee, a widow or widower’s pension and children pension under the Pension Scheme, and the assurance benefit under the Insurance Scheme. An answer that treats the three as one pot will misstate at least two of the three entitlements.

Inside the member’s account: membership, nomination and exit

Membership of the Provident Fund Scheme begins by operation of the scheme, not by enrolment at the member’s initiative. Every employee in a covered establishment who is not an excluded employee becomes a member of the Fund from the day the scheme begins to apply to the establishment or the day the employee joins it, whichever is later, and the employer’s duty to deduct and remit runs from that day. The scheme’s membership is thus a status imposed by the coverage of the establishment, which is why disputes about a missing account are usually disputes about coverage or about the wage base, and rarely about whether the employee agreed to join.

Nomination is the account’s succession machinery. A member nominates the person or persons to receive the accumulation on death, and where the member has a family the nomination must be in favour of a member of the family as the scheme defines it; a nomination made in favour of an outsider while the member has a family is invalid to that extent, and a member who acquires a family after nominating must make a fresh nomination. The nomination governs the provident fund, and separate nomination operates for the pension and insurance benefits, a separation that produces real disputes when a nomination made for one benefit is assumed to cover the others. On death in service the three schemes pay three different payees by three different rules: the provident fund balance to the nominee, the pension to the widow or widower and children in the scheme’s order, and the assurance benefit to the nominee or family entitled under the Insurance Scheme.

Exit from the account takes more forms than retirement. The scheme permits final withdrawal on retirement at the age the scheme fixes, on retirement on account of permanent and total incapacity, on migration from India for settlement abroad, on retrenchment, and on leaving service followed by a period of unemployment that the scheme specifies, reduced withdrawals being available earlier on several of these events. During service, non-refundable advances are available for purposes the scheme enumerates, which include medical treatment of the member or family, marriage and post-matriculation education in the family, purchase of a dwelling house or site, construction or repair of a house, and repayment of a housing loan. The advances are charged against the member’s own balance, so their cost is measured in the interest the withdrawn sum no longer earns. An account to which no contribution flows for a sustained period becomes what the organisation calls an inoperative account; the scheme’s rules govern the interest such accounts earn, and the accumulations remain payable to the member or the member’s nominees, since inoperation is an administrative state and not a forfeiture. Interest, when credited, is computed on the monthly running balances in the account, which is why a withdrawal taken early in a year costs more in forgone interest than the same withdrawal taken late.

The advance provisions deserve a word of their own, because they reveal the scheme’s second purpose. A provident fund designed purely for retirement would forbid every withdrawal before superannuation; the Provident Fund Scheme instead permits advances and partial withdrawals for a stated list of life events, marriage in the family, education of children, medical treatment, the purchase or construction of a house, the repayment of a housing loan, and specified hardships, each with its own service condition, ceiling and proof. The list is a piece of social policy embedded in a savings statute: the fund is made to serve as the family’s lender of first resort at the moments when a household without savings would borrow at moneylender or card rates, and the housing advances in particular have made the fund one of the quiet instruments of working-class house ownership in India. The discipline of the design is that the advance is the member’s own money lent to the member, repayable in the marriage and housing cases not at all, since it is a withdrawal, and the price is paid silently in the reduced balance that compounds thereafter. A member who draws repeatedly for consumption arrives at retirement with the pension intact and the lump sum spent, which is the scheme’s standing compromise between the emergency of the present and the poverty of old age, struck in favour of the present more often than a pure retirement design would allow.

What the member owns: the legal character of the accumulation

The accumulation in a member’s account has a distinctive legal character that the Act constructs from three directions. Section 5 provides for the Fund to vest in and be administered by the Central Board: the money is not the employer’s, to be reclaimed on a change of mind, and it is not the Government’s, to be spent as revenue; it is held by a statutory board for the members. The member’s right against the fund is a statutory entitlement measured by the account, enforceable through the scheme’s claim machinery rather than by an ordinary suit for a debt, and it survives the employer’s failure: an employer who deducted the employee’s share and collapsed without depositing it leaves a claim that the priority provisions pursue against the estate, because the deduction was never the employer’s money in any sense the law recognises. The third direction is Section 9, under which the Fund is recognised for the purposes of the income tax legislation, the recognition that carries the tax treatment of contributions, interest and withdrawals. That treatment has itself been qualified in recent years: since April 2021 the interest attributable to an employee’s own contributions above a high annual threshold fixed by the Finance Act of that year is taxable, a change that touches only the largest accounts and leaves the architecture of recognition intact. The composite picture is of deferred wages converted, by statutory machinery, into a protected individual asset with a tax shelter, administered by a tripartite board, and ranking ahead of the employer’s creditors and the member’s own.

The contribution: how Section 6 of the EPF Act 1952 reached twelve per cent

Section 6 fixes the price of the system. In its present text it provides that the contribution paid by the employer to the Fund shall be ten per cent of the basic wages, dearness allowance and retaining allowance, if any, payable to each employee, whether employed directly or by or through a contractor, and that the employee’s contribution shall be equal to the employer’s. A proviso then empowers the Central Government, after such inquiry as it deems fit, to specify establishments or classes of establishments for which the words “twelve per cent” stand substituted for “ten per cent” at both places in the section. An employee who wishes to contribute more may do so; the section states in terms that the employer is under no obligation to match the excess. Three features of this text repay attention. The rate in the section is a floor with a notified ceiling built in, not a single figure. The base is a statutory composite, basic wages plus dearness allowance (which Explanation 1 deems to include the cash value of any food concession) plus retaining allowance, the allowance payable for keeping an employee’s services while the establishment is not working. And the obligation runs to contract labour exactly as to direct employees.

The figures in the section are the residue of a seventy-year climb. When the first scheme took effect on 1 November 1952, the contribution was one anna in the rupee, which is six and a quarter per cent, from each side. An enhanced rate of eight per cent was applied to larger establishments in the years that followed. The amending Act 33 of 1988 standardised the rate in the section at eight and one-third per cent with effect from 1 August 1988, and from 1990 an enhanced ten per cent applied to establishments employing fifty or more persons. In April 1997 a notification extended the ten per cent rate to nearly every notified class of industry, and with effect from 22 September 1997 the notified enhancement moved each band upward: establishments at eight and one-third per cent rose to ten, and establishments at ten rose to twelve. The amending Act 10 of 1998 then wrote the settled figures into the section itself, substituting “ten per cent” for “eight and one-third per cent” and “twelve per cent” for “ten per cent” in the proviso, with retrospective effect from 22 September 1997. The twelve per cent that payroll applies is therefore not the rate Parliament printed as the general rule; it is the enhancement, applied by notification to the general run of covered establishments, resting on the ten per cent the section names. A smaller list of classes, notified under the proviso, continues at ten per cent.

What the employer pays is not what the provident fund account receives, and the division is the second fact members consistently misremember. Of the employer’s twelve per cent, eight and one-third per cent of wages is diverted to the Pension Fund under Section 6A(2), subject to the wage ceiling: the diversion is computed on wages up to Rs 15,000 a month, so the largest monthly diversion is Rs 1,250. The balance of the employer’s contribution, three and two-third per cent where wages are at or below the ceiling, is credited to the provident fund account. The employee’s twelve per cent goes entirely to the provident fund account. Where an employer and employee have jointly elected contribution on wages above the ceiling, the excess of the employer’s contribution over the pension diversion flows to the provident fund account as well, so the account’s share rises with salary while the pension diversion stays capped. The Central Government adds its own contribution to the Pension Fund of 1.16 per cent of wages, a figure that has survived from the family pension era into the pension scheme. Beyond the two contributions, the employer pays administrative charges for the Provident Fund Scheme, fixed at one-half of one per cent of wages since 1 June 2018, and the insurance contribution of one-half of one per cent, together with a further sum toward the administration of the Insurance Scheme capped by Section 6C(4) at one-fourth of the insurance contribution. Section 12 closes the obvious escape: an employer may not, by reason only of the liability to contribute, reduce an employee’s wages directly or indirectly, or cut the total quantum of old age, pension, gratuity or life insurance benefits promised by the terms of employment. The section has acquired a modern setting in the cost-to-company pay structure, under which an employer states a single annual figure as the employee’s package and treats the employer’s provident fund contribution as a component of it. Whether such a structure offends Section 12 turns on the terms of the contract of employment: the section forbids a reduction of wages by reason of the liability, and a package agreed at the outset, in which the contribution is one priced component among others, presents a different question from a wage cut imposed on an existing salary when coverage begins. The payroll function should treat Section 12 as a drafting rule: the contribution must never appear as a deduction from a wage already promised, and the wage base on which it is computed is the Section 2(b) base whatever the package document calls its components.

The rhythm of the obligation is monthly and unforgiving of drift. The employer deducts the employee’s share from the wages of the month and must remit both shares, with the charges, within fifteen days of the close of the month to which they relate. A worked example shows the price of a single quarter’s neglect. On a wage bill of Rs 10 lakh a month, the monthly remittance of contributions alone is Rs 2.4 lakh. Delayed by five months, the arrears of Rs 12 lakh attract interest under Section 7Q at twelve per cent per annum for the period of delay, and damages under Section 14B at the graded scheme rates, which for a delay beyond four months stand at fifteen per cent per annum and beyond six months at twenty-five per cent per annum, subject in each case to the statutory ceiling of the arrears themselves and to the authority’s duty to weigh the circumstances. The combined price of the delay approaches a fifth of the sum withheld, before the criminal exposure for the deducted share is considered, which is why provident fund default is among the most expensive forms of short-term borrowing available to an Indian employer.

The table below traces one month of wages of Rs 100, at or below the ceiling, through the system. It is the findable centre of this article, because no payslip in the country prints it.

Flow	Rate per cent of wages	Destination	Authority
Employee’s contribution	12.00	Provident fund account of the member	Section 6
Employer’s contribution diverted	8.33	Pension Fund	Section 6A(2)
Employer’s contribution retained	3.67	Provident fund account of the member	Section 6
Central Government contribution	1.16	Pension Fund	Section 6A(2)(d)
Employer’s insurance contribution	0.50	Deposit-linked Insurance Fund	Section 6C(2)
Employer’s administrative charges	0.50	Provident Fund Scheme administration	Section 6 read with the Scheme
Employer’s insurance administration sum	Up to one-fourth of the insurance contribution	Insurance Scheme administration	Section 6C(4)

The wage ceiling: a scheme figure that decides the pension

The figure of Rs 15,000 a month does more work in this Act than any section number, and it is routinely misdescribed as a figure in the Act. It is not. The Act nowhere fixes a wage ceiling. The ceiling lives in the schemes: most visibly in the definition of “excluded employee” in paragraph 2(f) of the Provident Fund Scheme, and in the corresponding provisions of the Pension Scheme that cap pensionable salary. The distinction matters for method. A ceiling in the Act could be moved only by Parliament; a ceiling in a scheme is moved by notification, and the history of the figure is a history of notifications: Rs 300 a month when the first scheme began in November 1952, then Rs 500, Rs 1,000, Rs 1,600, Rs 2,500, Rs 3,500 and Rs 5,000 in turn, Rs 6,500 from 1 June 2001, and Rs 15,000 from 1 September 2014, raised by the amendment notifications of 22 August 2014, of which G.S.R. 608(E) amended the Provident Fund Scheme and G.S.R. 609(E) amended the Pension Scheme. The 2014 revision more than doubled the ceiling at a stroke, and no revision has followed it. The revision did not travel alone: the notification package of August 2014 also fixed a minimum pension of Rs 1,000 a month under the Pension Scheme and added a twenty per cent additional relief to the assurance benefit under the Insurance Scheme, so that the three schemes were each adjusted in the same Gazette season, a reminder that the ceiling is one dial in a panel and that governments have historically preferred to move the panel together or not at all.

The ceiling performs three distinct jobs, and argument about it goes wrong when they are merged. First, it is the membership line: an employee drawing wages above the ceiling on joining a covered establishment is an excluded employee, outside compulsory membership, although a joint election can bring the employee in. Second, it caps the pension diversion and therefore the pensionable salary: the Pension Fund receives its eight and one-third per cent on wages up to the ceiling and nothing beyond it, unless a valid option for pension on higher salary subsists. Third, it sets the minimum statutory contribution: an employer of a member at the ceiling owes contributions computed on Rs 15,000 even though the member’s actual wages may be lower in a short month, since the scheme reckons the obligation on wages payable. The practical consequence of the second job dominates the politics of the Act. Pension under the Pension Scheme is computed on pensionable salary, and for the ordinary member pensionable salary is capped at the ceiling, so a working life spent at wages several times the ceiling produces a pension calculated on Rs 15,000 a month.

The case for the ceiling, stated at its strongest, is a case about the character of the statute. The Act is a floor of social security, not a complete retirement plan: it compels the minimum that every covered worker must have, and leaves wages above the floor to private arrangement, voluntary election under the scheme, and other instruments. A ceiling confines the employer’s statutory cost to a sum that marginal establishments can bear, and raising it raises the price of formal employment at exactly the margin, the small establishment deciding whether to hire on the rolls, where formalisation is most fragile. The case against, also stated at its strongest, is arithmetic. A ceiling unrevised for more than a decade while nominal wages rise means the statutory promise covers a shrinking fraction of the wage it purports to insure; the pension computed on a salary frozen in 2014 prices is a very small sum in the prices of the year it is drawn; and the joint election machinery that is supposed to cure the gap depends on employer consent and paperwork that the litigation over higher pension has shown to be fragile. Both cases are honestly held. The verdict the record supports is narrow: the ceiling is a legitimate instrument of a floor statute, but a ceiling that is never revised converts the instrument into a slow repeal of the benefit, and the revision cycle is a policy choice for which the notifying government, and not the Act, is answerable.

Who fixes the interest credited to a member’s account each year?

The Central Government declares the rate each year, on the recommendation of the Central Board of Trustees, and interest is credited to members’ accounts under the Provident Fund Scheme on monthly running balances. The rate is a single figure for the whole fund, set after the Board has taken stock of the fund’s earnings.

Two members at two salaries: the ceiling at work

The ceiling’s operation is clearest in figures. Take first a member whose wages are exactly Rs 15,000 a month. The employee contributes Rs 1,800 and the employer contributes Rs 1,800, of which Rs 1,250, being eight and one-third per cent of the wage, is diverted to the Pension Fund and Rs 550 is credited to the provident fund account. The member’s account therefore grows by Rs 2,350 a month before interest, the Central Government adds Rs 174 to the Pension Fund, and the pension, when it comes, is computed on the full wage because the wage never exceeded the ceiling. Take next a member whose wages are Rs 40,000 a month, in an establishment where no joint election has been made. The statutory obligation is confined to the ceiling: contributions are computed on Rs 15,000, and the flows are identical to the first member’s, rupee for rupee, although the second member earns more than two and a half times the first member’s wage. The deduction on the payslip may be larger if the establishment contributes voluntarily on actual wages, but the statute compels nothing beyond the ceiling flows.

A joint election changes the second case in one direction only. If employer and employee elect contribution on the actual wage, the employee contributes Rs 4,800 and the employer Rs 4,800, but the pension diversion remains Rs 1,250, because the diversion is capped at the ceiling unless a higher pension option subsists; the remaining Rs 3,550 of the employer’s contribution is credited to the provident fund account. The account then grows by Rs 8,350 a month before interest, while the pension promise is unchanged. The asymmetry is the whole politics of the higher pension question in miniature: contributing on actual wages is easy and merely enlarges a savings account, while pension on actual wages requires the option, the joint exercise, the transfer of past accumulations between the funds, and an actuarial liability the Pension Fund must carry for life.

Basic wages: the EPF Act 1952 definition that decides the bill

Every contribution figure in this article is a percentage of a base, and the base is litigated more than any other element of the Act. Section 2(b) defines “basic wages” as all emoluments which are earned by an employee while on duty, or on leave or on holidays with wages, in accordance with the terms of the contract of employment and which are paid or payable in cash to the employee. The definition then excludes three groups: the cash value of any food concession; any dearness allowance, house rent allowance, overtime allowance, bonus, commission or any other similar allowance; and any presents made by the employer. Section 6 then computes the contribution on “basic wages, dearness allowance and retaining allowance”, so dearness allowance, excluded from the definition, is added back by the charging section. The contribution base in law is therefore basic wages in the Section 2(b) sense, plus dearness allowance including the cash value of a food concession, plus retaining allowance if any. A payroll that contributes on the figure its own software labels “basic” without testing that figure against Section 2(b) has not begun the statutory inquiry.

The courts gave the definition its working rule early. In Bridge and Roof Co. (India) Ltd. v. Union of India, AIR 1963 SC 1474, the Supreme Court asked what distinguishes an allowance that is part of basic wages from one that is not, and answered with a test of universality: an emolument that is universally, necessarily and ordinarily paid to all employees across the board, as part of the wage structure, is basic wages, while a payment linked to an incentive for extra work, or to a contingency that may not arise, is not. The test looks through labels to the structure of the pay packet. Its corollary occupied the courts for decades: an employer cannot shrink the contribution base by splitting a single wage into a small “basic” component and a tail of allowances that every employee in fact receives every month, because the allowances fail the exclusion precisely by being universal.

The Supreme Court restated the position in Regional Provident Fund Commissioner v. Vivekananda Vidyamandir, decided on 28 February 2019. The establishments before the Court paid special allowances that were not linked to any additional output or special work, and the question was whether those allowances formed part of basic wages. The Court held that they did: allowances paid universally to all employees, without connection to an incentive or to extra work beyond the ordinary terms of employment, fall within basic wages for contribution purposes, and the exclusions in Section 2(b) protect only payments of the specific kinds the clause names or genuinely similar contingent payments. The judgment consolidated the Bridge and Roof line rather than departing from it. Between the two decisions the rule for the payroll function is stable and can be stated in one sentence: the contribution base is the whole of the ordinarily universal wage, minus only the payments the definition expressly excludes, and the employer’s own nomenclature carries no authority against the statute. The successor definition in the Code on Wages, 2019 approaches the same mischief from the opposite direction, by capping the share of the total wage packet that may be loaded into excluded allowances; the wage definition question therefore survives the codes in a new form, and the Section 2(b) case law remains the authority for the years the Act governed and governs.

The changing establishment: transfers, common funds and accumulated balances

Establishments merge, split and change hands, and the Act anticipates each event because each is a classic occasion for accumulations to vanish. Section 15 governs the first transition of 1952 and its successors: where an establishment already ran a provident fund of its own when the scheme was applied to it, the accumulations standing to the credit of employees who became members were transferred to the statutory Fund and credited to their accounts, notwithstanding anything in the deed or instrument that established the private fund. The private fund’s terms could not be set up against the statute. Section 17B governs transfers of ownership: where an employer transfers an establishment, by agreement or by operation of law, to another person, the transferor and the transferee are jointly and severally liable to pay the contributions and other sums due up to the date of transfer, with the transferee’s liability limited to the value of the assets obtained by the transfer. The limitation protects a buyer who priced the acquisition on the assets, while the joint liability prevents the seller from escaping by the act of sale. Section 3 addresses the opposite configuration, establishments under different employers that share a common provident fund, and empowers the Central Government to direct that the Act apply to them as if the fund were one, so that a common fund cannot be used to fragment a single economic unit into uncovered pieces. Read with Section 2A’s rule on departments and branches, the pattern is constant: the Act follows the economic establishment through changes of legal dress, and treats the member’s accumulation as a charge on the enterprise rather than a favour of its momentary owner.

Determination, interest and damages: the price of default under the EPF Act 1952

The Act assumes a defaulting employer and builds its machinery accordingly. The front of the machinery is Section 7A, inserted in 1963 and recast in 1988. An authorised officer, who may be the Central Provident Fund Commissioner or any Additional, Deputy, Regional or Assistant Commissioner, may decide a dispute regarding the applicability of the Act to an establishment and may determine the amount due from an employer under the Act or any of the three schemes. For that inquiry the officer holds the powers of a court under the Code of Civil Procedure in enforcing attendance, examining on oath, requiring discovery and production of documents, receiving evidence on affidavit and issuing commissions, and the inquiry is deemed a judicial proceeding for the purposes of the Indian Penal Code. No order may be passed without giving the employer a reasonable opportunity of representing the case, and an ex parte order may be set aside on application within three months where the employer shows the notice was not duly served or that sufficient cause prevented appearance. Sections 7B and 7C add a review of a Section 7A order and a power to determine amounts that escaped assessment, and an appeal lies to the Employees’ Provident Funds Appellate Tribunal under Section 7I, with Section 7O requiring the appellant to deposit twenty-five per cent of the amount determined as a condition of the appeal being entertained, a deposit the Tribunal may waive or reduce. The Tribunal itself, constituted under Section 7D, is presided over by a person who is, has been, or is qualified to be, a Judge of a High Court or a District Judge, a qualification the amending Act 10 of 1998 widened to include the district judiciary. The forum map that results allocates each dispute to its decider: applicability and quantum to the Section 7A officer, review to the same office under Section 7B, escaped assessment to Section 7C, appeal on the determination to the Tribunal, and the constitutional writ jurisdiction of the High Courts over all of it, which is where the large questions, the wage definition and the pension amendment among them, were in fact fought. An employer who litigates a coverage dispute as though it began in a court misunderstands the design: the Act builds a self-contained ladder, and the courts intervene from above it rather than within it. The inquiry runs on records, and the record-keeping duty is part of the liability structure: the schemes oblige the employer to maintain the returns and registers open to the inspectors appointed under Section 13, and where an employer withholds or fails to maintain the records from which the workforce and the wage base are proved, the officer determines on the material available, and assessments built on the best material the employer left available are accepted on challenge, since a party cannot found an appeal on the absence of evidence it was bound to keep.

The inspectors appointed under Section 13 are the inquiry’s eyes. An inspector may, for the purposes of the Act, inquire into the correctness of information in a return, enter and search any establishment and require the production of books, registers and records, examine the employer, the agents and the persons found in charge, make copies of or take extracts from the records, and seize such records as appear necessary where an offence is suspected, with the powers in relation to search and seizure following the Code of Criminal Procedure. The inspection is not a formality layered over the assessment: it is the stage at which the workforce is counted, the wage registers are compared with the returns, and the discrepancy that a Section 7A notice will later quantify is first written down. Employers experience the machinery in sequence, return, inspection, inquiry, assessment, and each stage’s record becomes the next stage’s evidence, which is why an establishment that treats the return as a negotiating position discovers that the position has been recorded.

Delay is priced by two separate provisions that employers persistently conflate. Section 7Q imposes interest as a matter of course: the employer is liable to pay simple interest at twelve per cent per annum, or at such higher rate as the scheme may specify, on any amount due under the Act, from the date the amount became due until the date of actual payment. Interest is compensation for the time value of money that belonged to the fund, and it admits no discretion and no waiver. Section 14B is the penal limb. Where an employer defaults in paying a contribution, in transferring accumulations, or in paying charges, the Central Provident Fund Commissioner or an authorised officer may recover from the employer, by way of penalty, damages not exceeding the amount of the arrears, at the rates the schemes specify. The Provident Fund Scheme grades those rates by the length of the default, rising from five per cent per annum for a delay under two months to twenty-five per cent per annum for a delay beyond six months. Before damages are levied the employer must be given a reasonable opportunity of being heard, and the second proviso reserves to the Central Board a power to reduce or waive damages for a sick industrial company under a sanctioned rehabilitation scheme. Interest under Section 7Q and damages under Section 14B therefore answer different questions: the first asks how long the fund was kept out of its money, the second how blameworthy the keeping was, and a defaulting employer ordinarily pays both. The blameworthiness inquiry has a leading authority. In Organo Chemical Industries v. Union of India (1979), the Supreme Court characterised the damages under Section 14B as a penal levy imposed to penalise the employer and to compensate the fund, and held that the levying authority must apply its mind to the circumstances of the default, including its frequency, duration and the reasons offered, rather than impose the maximum mechanically. The judgment is the reason a Section 14B order must be a speaking order: the graded rates in the scheme frame the discretion, they do not discharge it, and an order that does not engage with the employer’s explanation is vulnerable in appeal and in the writ court. The civil and criminal limbs run on separate tracks: a Section 7A determination, interest and damages proceed on the footing of liability to pay, and neither requires nor awaits a prosecution, while a prosecution under Section 14 turns on the ingredients of the offence and the sanction of the complaint under Section 14AC, so an employer may be assessed, charged interest and damages, and prosecuted for the same default without any limb barring the others, the amounts recovered in the civil process being credited to the funds and not to the State.

Behind the civil machinery stands a criminal provision aimed at the one default the Act treats as a species of misappropriation. Section 14(1A) punishes default in complying with Section 6, and a default in paying the employees’ contribution that the employer has deducted from wages carries a minimum sentence: imprisonment of not less than one year and a fine of ten thousand rupees, within a maximum of three years. The reasoning is visible on the face of the scheme. The employer’s own contribution is a debt; the employee’s contribution, once deducted, is the employee’s money in the employer’s hands, and the Act prices its retention accordingly. Section 14C empowers the convicting court to order payment of the contribution in addition to the punishment. The penalty chapter is completed by Section 14A, which fastens liability for offences by companies on the persons in charge of and responsible for the conduct of the business, by Section 14AB, which makes the graver contribution offences cognizable, and by Section 14AC, which channels prosecution through a complaint by the authorised officer. Prosecution under the Act is therefore an instrument held in reserve behind the assessment and recovery machinery: the ordinary defaulter meets interest, damages and certificate recovery, and the criminal process is reserved, in the main, for the deducted-but-not-deposited case and for false statements made to avoid payment, which Section 14(1) punishes with imprisonment of up to one year or a fine of five thousand rupees or both.

Recovery, attachment and priority: why provident fund dues are hard to defeat

A determined sum is only as good as its recovery, and Section 8 gives the determination teeth. Amounts due from an employer may be recovered as arrears of land revenue, and Sections 8B to 8G construct a certificate machinery of the employer’s own: the authorised officer issues a certificate to a Recovery Officer, who proceeds against the defaulter by attachment and sale of movable or immovable property, by arrest and detention, or by appointing a receiver for the management of the property, with Section 8G applying the attachment and sale provisions of the Income-tax Act, 1961 to the process. The certificate cannot be disputed by the defaulter on the merits in the recovery forum; the challenge belongs to the Section 7A appeal. Section 8F adds further modes, among them recovery from a person who holds or may later hold money for or on account of the defaulter, on notice from the authorised officer, so that a debtor of the establishment can be required to pay the fund instead of the establishment. Section 8A completes the circuit for contract labour by letting the principal employer recover sums paid for a contractor’s workers from the contractor. The employer in default therefore faces a revenue recovery of the kind the tax statutes use, without the fund having to file a suit.

Protection runs in the opposite direction as well. Section 10(1) provides that the amount standing to the credit of a member in the Fund is not capable of being assigned or charged, and is not liable to attachment under any decree or order of any court in respect of any debt or liability incurred by the member. Section 10(2) carries the protection through death: a sum payable to a nominee on a member’s death vests in the nominee and is free from any debt or liability incurred by the deceased member or the nominee before the death. The accumulation is walled off from the member’s own creditors in life, and from the deceased’s estate debts after death, because the statute treats the fund as deferred wages held for subsistence, not as an asset available to the credit market.

The priority provisions address the employer’s insolvency, and they are layered with unusual care. Section 11(1) provides that where an employer is adjudicated insolvent or a company is ordered to be wound up, the contributions, Section 14B damages, transferable accumulations and charges accrued before the order rank among the debts payable in priority to all other debts in the distribution of the property or assets. Section 11(2), inserted in 1973, goes further and speaks in charge language: any amount due from an employer, whether the employee’s contribution deducted from wages or the employer’s own, “shall be deemed to be the first charge on the assets of the establishment, and shall, notwithstanding anything contained in any other law for the time being in force, be paid in priority to all other debts”. The modern insolvency framework reinforces the position from the other side: under the Insolvency and Bankruptcy Code, 2016, the liquidation estate itself excludes all sums due to any workman or employee from the provident fund, the pension fund and the gratuity fund, so those sums never enter the pool over which creditors contend, a rule the insolvency and bankruptcy code guide examines within the Code’s distribution scheme. Seed FAQ territory treats the insolvency question directly; the structural point for the pillar account is that the protection is triple: the money in the Fund is exempt from attachment, dues not yet paid carry a statutory first charge, and sums attributable to the funds are kept out of a liquidation estate altogether. The force of the first charge was tested against secured creditors under the companies legislation in Employees Provident Fund Commissioner v. Official Liquidator of Esskay Pharmaceuticals Ltd., decided in 2011, where the Supreme Court held that provident fund dues, by virtue of Section 11(2), rank ahead of the claims of secured creditors in the distribution of a wound-up company’s assets, the statutory charge being treated as a first charge in substance and not a mere priority among unsecured debts. The judgment matters beyond its facts because it fixes the meaning of the deeming language: the amount due is not simply paid early, it is charged on the assets themselves, and a lender taking security over an establishment’s assets takes it subject to a charge the borrower may not yet have quantified. For the same reason the charge binds in the hands of a transferee of the establishment, subject to the limit Section 17B places on a transferee’s liability for the period before the transfer.

The practical consequence for credit is worth stating plainly, because it is the point at which the Act touches persons who never employ anyone. A bank or debenture trustee that takes security over a factory takes it behind a provident fund charge whose size depends on the borrower’s payroll discipline, a discipline the lender cannot observe from the balance sheet, since undetected default leaves no trace in audited accounts until a Section 7A order quantifies it. Prudent lending against establishment assets therefore treats provident fund compliance as a diligence item in its own right, verified against the organisation’s records rather than the borrower’s representations, and the standard security documents in Indian industrial lending have carried provident fund compliance undertakings for decades for exactly this reason. The first charge is, in operation, a rule of commercial law no less than of labour law.

Can an employer run its own provident fund instead of the statutory fund?

Yes, by exemption under Section 17, where the appropriate Government is satisfied that the establishment’s own provident fund rules and the benefits under them are, on the whole, not less favourable than the statutory rules and benefits. The exempted establishment must set up a board of trustees and submit to inspection; Sections 6, 7A, 8 and 14B continue to apply.

Exemption in practice: the trust sector inside the EPF Act 1952

The exempted sector is large enough to be a system within the system, and its working rules deserve a fuller statement than the grant of exemption alone. An exempted establishment’s trust is governed by rules approved under the Act, administered by a board of trustees on which the employees are represented, and the trust must invest the accumulations in accordance with the pattern prescribed for the statutory Fund, file periodical returns with the regional office, and submit to inspection, for which the establishment pays inspection charges under Section 17(3) in place of the administrative charges. The exemption’s condition of benefits not less favourable is a continuing condition and not a threshold crossed once: a trust whose declared interest falls short of the rate declared for the statutory Fund owes its members the difference in substance, and the shortfall question is the standing item of friction between exempted establishments and the organisation. Cancellation under Section 17(4) follows where the conditions fail, and on cancellation the accumulations transfer to the statutory Fund and the members become members of the statutory scheme, with their service and balances carried over. Members of exempted trusts hold one protection the statutory members do not need and one exposure they do not share: the protection is a trust deed and a board on which their representatives sit, and the exposure is that the quality of their retirement provision depends in part on the governance of a single employer’s fund, which is why the successor scheme of 2026 for the exempted sector, framed under the Code, tightened the governance and reporting conditions on exempted funds in the reporting of the reform. The sector’s existence is the Act’s concession to institutional variety; its supervision is the price of the concession.

The schemes as subordinate legislation: why the detail sits where it sits

A reader comparing the Act with the schemes will notice the division of labour and should understand its rationale, because it recurs across Indian welfare legislation. Parliament enacted the obligations that could not be delegated without weakening them: the duty to contribute, the offence of retaining the deduction, the first charge, the protection from attachment and the determination machinery. The matters consigned to the schemes are those that must move at the speed of administration: rates within the statutory bands, ceilings, benefit formulae, claim procedures and forms. The pension demand itself illustrates the political logic: a retirement pension under the Act was pressed in Parliament and in the tripartite forums through the nineteen eighties, and when it came in 1995 it came as a scheme under an enabling section inserted in 1996, not as a new Act, because the scheme route allowed the benefit to be designed, valued and adjusted by notification under the laying procedure. The laying procedure is the formal check on the delegation: every scheme and every amendment is laid before Parliament, which may modify or annul it without prejudice to what has been done. In practice the check that bites is judicial rather than parliamentary, and the higher pension litigation, in which a scheme amendment was tested against the parent Act and the Constitution, is the working example of the control the design intends.

The member’s side of enforcement: remedies when the employer does not pay

The enforcement sections are written around the employer, but the member is not a bystander in them. A shortfall surfaces in the member’s account as a missing credit, and the scheme’s passbook and the universal account number’s online record make the shortfall visible to the person it harms, which was not the position in the era of the paper ledger. The member’s practical remedies run in a sequence. The first is the organisation’s grievance machinery, which registers a complaint against the establishment and routes it to the regional office holding jurisdiction, and a complaint by employees is itself material on which a coverage or quantum inquiry can be founded, since Section 7A proceedings may be set in motion on information available to the officer and are not confined to the employer’s returns. The second is the inquiry itself: in a Section 7A proceeding the employees’ evidence on the wage base and the workforce is frequently the decisive material, and the officer’s powers to summon documents and witnesses exist for exactly the case where the employer’s records are incomplete or arranged. The third is the criminal process for the deducted share, which the Act reserves to a complaint by the authorised officer under Section 14AC, so the member’s route to prosecution runs through persuading the organisation rather than filing a private complaint. The design is deliberate: enforcement is institutional, and the member’s power lies in visibility, in evidence and in the trigger, not in conducting the recovery. A member who discovers a default therefore accomplishes most by placing the wage evidence, payslips, appointment terms and the account record, before the regional office in a form the inquiry can use, and by preserving the same record for the claim that follows recovery.

The record a member keeps decides the member’s own claims as well. Pension turns on the service record and the wage record together; the higher pension litigation turned, in thousands of individual cases, on whether a joint option form said to have been filed years earlier could be proved, and the members whose claims survived scrutiny were those whose employers’ validations matched a paper trail. The habit of keeping the option forms, the joint declarations, the wage slips and the correspondence is unglamorous and decisive, and a member assembling such a file can keep it in any durable form, including a set of scanned papers read through a presentation viewer for the assembled claim record, so long as the sequence of documents can be produced complete when a field office asks for it.

Two governments, one fund: the appropriate Government question

Section 2(a) divides constitutional responsibility for the Act between the two levels of the Indian federation. In relation to an establishment belonging to or under the control of the Central Government, or connected with a railway company, a major port, a mine or an oilfield or a controlled industry, or having departments or branches in more than one State, the appropriate Government is the Central Government; in relation to any other establishment it is the State Government. The division governs such matters as the appointment of inspectors and the grant of exemptions, which is why an exemption notification for a multi-state establishment issues from the Centre while a single-state establishment looks to its State. Administration of the funds themselves does not divide: the Central Board and the organisation administer the schemes across the whole field, so an employer may owe its exemption to one government and its monthly remittance to a single national machinery. The arrangement is characteristic of the Act’s design, federal in its authorities and unitary in its money.

Administration: a board, an organisation and a number

The funds vest in and are administered by the Central Board of Trustees under Section 5A. The Board is tripartite in composition: representatives of the Central Government and the State Governments sit with representatives of employers and of employees, and the Central Provident Fund Commissioner is its chief executive. Section 5B provides for State Boards, and Section 5C makes each board a body corporate. Beneath the Board, the Employees’ Provident Fund Organisation is the executive machinery: commissioners and inspectors appointed under Sections 5D and 13, regional and sub-regional offices that decide coverage questions, conduct Section 7A inquiries, run recovery and settle claims. Inspectors may enter and search premises, require production of accounts, books and registers, examine persons in charge and seize documents where an offence is suspected, powers that make the coverage inquiry an evidence-gathering exercise rather than a correspondence.

For the member, the administrative fact that matters is the Universal Account Number. The number, introduced in 2014, is a permanent identity for the member to which each employment’s member account is linked, so that changing employers adds an account under the same number rather than orphaning the earlier accumulation. Section 17A supplies the statutory counterpart: on a change of employment to another covered establishment, a member’s accumulations are transferred to the account with the new employer. The number converted portability from a paper transfer pursued by the departing employee into a property of the identity itself, and it is the reason the withdrawal question on changing jobs is ordinarily a question about pension service records rather than about locating the money.

The organisation’s hold on employers is exercised through a monthly discipline. Contributions for a month’s wages must be remitted within the time the Provident Fund Scheme fixes, fifteen days from the close of the month, accompanied by the electronic return through which each member’s share is posted to the member’s account. The return, and not the payment alone, is the compliance act: a remittance without its allocation data leaves money in the fund that belongs to no identified account, and the Section 7A inquiry treats the employer’s own returns as the primary evidence of the wage base. Around that monthly cycle the Act arranges the annual rhythm of interest declaration, the scheme’s returns, and the inspection power, so that an establishment’s provident fund life is a sequence of dated obligations rather than a standing intention.

The money itself is invested under direction. The three funds are invested in accordance with the pattern of investment notified by the Central Government, in government securities, other approved securities and instruments within the proportions the pattern fixes; since 2015 the pattern has permitted a portion of incremental accretions to be invested in equity through exchange traded funds, which is the fund’s only exposure to market risk on members’ behalf. The interest a member earns is therefore not a market return on an individual portfolio but a declared rate struck from the earnings of a pooled, predominantly debt portfolio, and debate about the rate is in substance debate about the pattern and about the surplus the fund carries.

The declared interest rate deserves a section of its own in any honest account, because it is the number by which most members judge the fund. The rate is not earned account by account. The funds are invested as pools under the notified pattern, the earnings of a year are struck for the pool, and the Central Board recommends a rate of interest for the year which the Central Government approves and declares, whereupon the rate is credited uniformly to every account on its monthly balances. The mechanism has two consequences that public argument often misses. A rate can exceed the year’s earnings only by drawing on accumulated surplus, and can fall short of them only by building surplus up, so the declaration is in part a decision about inter-year smoothing, transferring income between cohorts of members. And the rate is bounded by the investment pattern: a portfolio confined in the main to government and approved securities cannot yield what an equity portfolio yields in a rising market, which is the entire substance of the recurring debate over the equity share of the fund. The entry of a portion of incremental accretions into exchange traded funds from 2015 was the first breach in the debt-only pattern, and argument over its ceiling, between those who regard members’ savings as capital to be grown and those who regard them as wages to be preserved, is the investment form of the same argument the wage ceiling conducts in contribution form. The Act itself is neutral in the debate; it builds the pool, the pattern and the declaration machinery, and leaves the proportions to policy.

Exempted establishments stand inside this administration on altered terms. An establishment exempted under Section 17 runs its provident fund through its own board of trustees, which must maintain detailed accounts for each employee, submit returns to the Regional Provident Fund Commissioner, invest the monies as the Central Government directs, and transfer accounts when employees move. In exchange for the scheme’s provisions it pays inspection charges in place of administrative charges, and the exemption can be cancelled for breach of its conditions. Sections 6, 7A, 8 and 14B apply to the exempted employer in addition to the exemption conditions, so the rate, the determination power, recovery and damages follow the money into the private trust; what the exemption changes is the custodian, not the obligation.

The family pension ancestor: what the EPF Act 1952 scheme of 1971 did, and what 1995 changed

The pension did not arrive in 1995 from nothing. The Employees’ Family Pension Scheme, 1971 was built on a smaller promise: a monthly family pension to the widow or widower and minor children of a member who died in service, funded by diverting one and two-third per cent of wages out of the contributions otherwise payable to the provident fund, with the Central Government contributing a further 1.16 per cent of the wages. There was no retirement pension at all. A member who lived to retire received the provident fund and nothing monthly; the scheme insured only the family’s income against the member’s early death. The diversion carried a grievance that the trade union movement pressed for two decades: the pension slice reduced the lump sum of every member, including the great majority whose families never drew the family pension, in order to fund a benefit for a minority.

The Employees’ Pension Scheme, 1995 answered the grievance by enlarging the promise rather than by removing the diversion. The family pension’s funding stream was redirected and increased, the employee’s diversion was ended so that the employee’s whole twelve per cent returned to the provident fund account, the employer’s diversion rose to eight and one-third per cent, and the benefit became a retirement pension first and a family pension second, with the ceased scheme’s beneficiaries protected at not less than their earlier entitlement by Section 6A(3). The 1.16 per cent government contribution survived from the 1971 design into the new scheme, which is why that oddly precise figure appears in the funding of a scheme framed a quarter of a century later. The substitution was retrospective in its commencement, taking effect from 16 November 1995 under an amending Act passed in 1996, and members of the ceased scheme were carried into the new one with their past service counted, on the scheme’s terms, toward pensionable service. The 1995 design is therefore best read as a renegotiation of 1971: the same social purpose, a wider benefit, a heavier charge on the employer’s share alone, and a funding gap between the diversion and the promise that the actuarial valuations have measured ever since.

The pension promise in detail: the EPF Act 1952 scheme of benefits

Section 6A(1) names the benefits the Pension Scheme exists to provide, and the scheme’s catalogue is wider than the retirement pension that dominates discussion. The first head is superannuation pension, payable on the member’s retirement at the age of fifty-eight. Alongside it stand retiring pension for a member who leaves service before superannuation after the qualifying service the scheme requires, and permanent total disablement pension for a member disabled in service so as to be unfit for the employment the member was engaged in. The qualifying service for a monthly pension is ten years of eligible service; a member who leaves with less is entitled instead to a withdrawal benefit computed on the service rendered, or to a scheme certificate that preserves the service record so that a later period of membership can be aggregated with it. The scheme certificate is the pension system’s answer to the mobile worker: it carries the service, rather than the money, across employers.

Early pension is available from the age of fifty, at a pension reduced for the years by which the retirement precedes fifty-eight at the rate the scheme prescribes, and deferment of the pension beyond fifty-eight, up to the age of sixty, is rewarded with an augmentation. On the member’s death, the family heads of the scheme take over: a widow or widower’s pension for the spouse, a children pension for up to two children at a time, fixed by the scheme as a fraction of the widow pension for each child, and an orphan pension at a higher fraction where both parents are dead. A member without a family may nominate a person to receive a nominee pension on death. The minimum monthly pension under the scheme stands at Rs 1,000, fixed by the amendments of September 2014, a floor sustained as a matter of policy alongside the formula pension it supplements.

Membership of the Pension Scheme follows membership of the provident fund, with one historical seam. Employees who were members of the Family Pension Scheme and of the provident fund when the Pension Scheme commenced on 16 November 1995 were carried into the new scheme with their past service, and every employee who became a provident fund member thereafter, and who was not an excluded employee by reason of wages above the ceiling, became a pension scheme member as a term of the provident fund membership; pension is not an optional extra that a covered employee may decline while keeping the fund. The seam shows at the ceiling: an employee who joins on wages above Rs 15,000 a month is outside the pension scheme unless the joint election route is used, and an employee whose wages cross the ceiling in service remains a member, with the diversion and the pensionable salary capped, which is why the scheme’s population divides between those whose entire wage is pensionable and those for whom the pension is a ceiling-priced benefit riding on a larger salary.

The formula itself has already been stated: pensionable salary multiplied by pensionable service, divided by seventy. Three features of the formula deserve emphasis because they allocate the scheme’s generosity. The service counted is pensionable service, which runs from the member’s joining the Pension Fund and takes in past service under the ceased family pension scheme on the terms of the transition. The salary counted is not the last salary but an average across the sixty months before exit, a basis substituted by the 2014 amendment for the twelve-month average that preceded it, which damps the effect of a final promotion on the pension. And the divisor converts the product into a lifetime monthly payment whose capital value depends on survival, which is the point at which the scheme’s actuarial character enters: the same formula yields a modest total for a member who dies at sixty and a large one for a member who lives to ninety, and the fund, not the member’s estate, carries that difference.

Higher pension: the option, the judgment and the unsettled ledger

The pension is a formula, and the formula’s inputs decide its size. Monthly pension under the Pension Scheme is pensionable salary multiplied by pensionable service and divided by seventy, with a weightage of two years of service for a member who retires on completing twenty years of pensionable service. Pensionable salary is the average of the pay drawn in the sixty months preceding exit, and it is capped at the wage ceiling for every member who has not validly opted for pension on higher salary. The Pension Scheme as framed in 1995 carried an option in paragraph 11(3): a member, with the employer’s agreement, could have pension computed on actual salary, with the employer’s contribution diverted on that higher base. The option was the bridge between a ceiling-capped scheme and salaries that ran far above the ceiling, and its exercise, non-exercise and paperwork became the subject of a decade of litigation. Its social setting should be understood. In a large number of establishments, including many public sector undertakings and the exempted trusts, contributions had in practice been remitted on actual wages for years, in some because the pay structures predated the ceiling’s bite and in others under joint elections recorded in the trust’s or the employer’s papers, and the members of those establishments retired expecting, and in the years before the 2014 amendment frequently received, pension computed with reference to the salary on which the money had actually been paid. The 2014 amendment’s insistence on a fresh option within a fixed window, and the organisation’s refusal of options said to have been exercised outside the procedures it recognised, converted that expectation into the dispute: the legal question of the amendment’s validity sat on top of an administrative question about what counted as an exercise of the option, and the Supreme Court’s judgment allocated the two questions to different deciders, validity to itself and the individual exercises to the validation process that followed.

The first judicial settlement came in R.C. Gupta v. Regional Provident Fund Commissioner (2016), where the Supreme Court struck down a cut-off date by which the organisation had refused options exercised after the date, holding that the scheme’s option could not be defeated by a deadline the scheme itself did not contain. The Employees’ Pension (Amendment) Scheme, 2014, notified as G.S.R. 609(E) on 22 August 2014, had meanwhile rebuilt the option on new terms from 1 September 2014: the ceiling rose to Rs 15,000, pensionable salary moved to a sixty-month average, members drawing above the ceiling were required to make a fresh option within six months to remain in the scheme on higher salary, and those members were required to contribute an additional 1.16 per cent of the salary above the ceiling into the Pension Fund. Three High Courts struck the amendment down. The controversy that reached the Supreme Court was therefore double: whether the 2014 terms were valid at all, and what became of members who had served under the earlier option regime.

The Supreme Court answered on 4 November 2022 in Employees Provident Fund Organisation v. Sunil Kumar B. The Court upheld the 2014 amendment, holding that the classification it drew by salary and its conditions were not whimsical or arbitrary, and that the amendment operated prospectively. It read down one condition: the requirement that a member contribute an additional 1.16 per cent of the salary above the ceiling was held ultra vires the 1952 Act, since neither Section 6 nor Section 6A authorises a contribution out of the member’s wages to the Pension Fund, and the Court suspended the operation of that holding for six months to allow the scheme to be adjusted. Members who were in service on 1 September 2014 and had not exercised the higher pension option were given a further period of four months to exercise it jointly with their employers, and the facility was extended to members of exempted establishments. The judgment thereby settled the law and opened an administrative exercise: joint options had to be invited, validated against service records, and priced, because a pension on actual salary implied diverting to the Pension Fund sums that had for years been credited to individual provident fund accounts, with interest adjustments between the two funds.

Implementation ran through 2023 and beyond, and its record should be stated with attribution rather than characterised. The Press Information Bureau release of 3 May 2023 recorded the Government’s chosen route for the 1.16 per cent: rather than take the sum from members, the additional contribution would be drawn from within the employer’s overall twelve per cent, so that in a higher pension case the Pension Fund receives nine and forty-nine hundredths per cent of actual wages and the provident fund account the remaining two and fifty-one hundredths per cent, a reallocation the release described as retrospective in line with the judgment. The organisation invited joint options through an online facility, employers validated applications against wage records, and field offices issued demand notices for the differential amounts transferable between funds. Trade union trustees on the Central Board, among them Sunkari Mallesham of the Bharatiya Mazdoor Sangh, pressed publicly for an extraordinary meeting of the Board to work through the order’s implementation. The two positions in the underlying dispute retain their force after the judgment. The members’ case is that a pension computed on Rs 15,000 a month returns a small fraction of the last drawn salary to workers whose employers and who themselves contributed for decades, in many establishments on actual wages, and that the option was the scheme’s own promise. The organisation’s actuarial case, placed in its materials before the Court, is that the Pension Fund carries a large actuarial deficit, that each admission to pension on actual salary enlarges a defined benefit liability that the past diversions, computed at the ceiling, never funded, and that generosity to one cohort is paid for by the pool that serves every other member. The judgment holds the legal line between the two: the option survives for those the Court identified, on the scheme’s terms, and the funding gap is left where the statute leaves it, with the Government and the fund. Members of exempted establishments occupied a distinct corner of the exercise. Their accumulations sit in private trusts rather than in the statutory Fund, so a higher pension option required the trust and the Pension Fund to settle the transfer of the pension share of past accumulations between them, with the employer’s records as the evidence; the Supreme Court’s extension of the facility to exempted trust members settled their eligibility, and the mechanics of valuation and transfer were left to the organisation’s circulars, which is one reason the disposal of applications proceeded unevenly across regions, a state of affairs reported in the organisation’s own updates and in the press coverage of the facility rather than admitted in any single document.

Why did the pension provisions of the Code move in May 2023, years before the Code commenced?

Because the Supreme Court’s judgment of 4 November 2022 was time-bound. The Court suspended the invalidity of the 1.16 per cent condition for six months to allow the scheme to be adjusted, and the adjustment, drawing the amount from within the employer’s contribution, was carried into effect by the notifications of 3 May 2023.

A comparison fixes what kind of promise the pension is. Under the National Pension System governed by the PFRDA Act, 2013, examined in the article on the PFRDA Act 2013 and pension reform, the retirement benefit is an account balance exposed to markets, and the statute promises a framework, not an amount. The Pension Scheme under this Act promises an amount computed by formula and funded by a capped diversion, which is why its controversies are actuarial and legislative rather than commercial: every improvement in the promise is a charge on a pool, and every enlargement of the pool is a charge on wages or on the exchequer.

The insurance layer in detail: the EPF Act 1952 benefit nobody prices

The Deposit-linked Insurance Scheme is the only one of the three schemes whose cost falls on the employer alone and whose benefit the member never sees priced anywhere. Its design repays a closer statement. The employer’s contribution of one-half of one per cent of wages is not credited to any account; it is a premium paid into a pooled fund, and the benefit on a member’s death in service, the assurance benefit, is computed under the Insurance Scheme from two variables, the member’s wages and the average balance in the member’s provident fund account over the period the scheme specifies, within a minimum and a maximum that the scheme fixes and that have been revised upward by notification at intervals. The linkage to the balance is the scheme’s signature. A conventional group insurance pays by salary band; this benefit grows with the member’s own accumulation, so the longest-serving member, whose family is likeliest to have depended on the wage longest, draws the largest assurance, and the benefit operates in part as a return on the savings behaviour the rest of the Act compels. The linkage has a cost in comprehension: because the amount depends on the balance at death, no payslip and no scheme summary can tell a member in advance what the family will receive, and the benefit is discovered by most families only when it is claimed. The claim itself travels with the provident fund claim on death, and the same nomination and family rules govern the payee, which is why the organisation settles the three death benefits in a single proceeding even though three funds pay them.

Set against the other two schemes, the insurance layer also exposes the Act’s theory of risk. The provident fund refuses risk altogether and returns the balance. The pension scheme pools the risk of long life. The insurance scheme pools the risk of early death. A single contribution stream from a single wage thus buys the three elementary protections of a working life, survival to retirement with savings, survival past retirement with income, and non-survival with a payment to the family, and the completeness of the trio, achieved by 1976, is the strongest structural argument for the Act’s design and the reason its repeal-and-replacement has always faced a higher burden of proof than the repeal of a single-purpose statute.

The repeal of the EPF Act 1952 that was notified for everyone else

The commencement story is the article’s most valuable content, and it has to be told in sequence, because each step reverses the impression the step before it creates. The Code on Social Security, 2020 (Act 36 of 2020) contains, in Section 164, a repeal and savings provision. Sub-section (1) lists nine enactments to be repealed: the Employees’ Compensation Act, 1923, the Employees’ State Insurance Act, 1948, this Act of 1952 as the third item in the list, the Employment Exchanges (Compulsory Notification of Vacancies) Act, 1959, the Maternity Benefit Act, 1961, the Payment of Gratuity Act, 1972, the Cine-Workers Welfare Fund Act, 1981, the Building and Other Construction Workers’ Welfare Cess Act, 1996, and the Unorganised Workers’ Social Security Act, 2008. A reader of the Code’s text alone would conclude that all nine stand repealed. The series’ labour cluster articles, in describing the nine enactments as subsumed by the Code, described exactly that: the Code’s text. Repeal under the Code, however, is not self-executing. Section 1(3) of the Code lets the Central Government appoint different days for different provisions, and Section 164 operates only as and when it is brought into force, and for the enactments it is brought into force against. The social security code explainer and the account of the labour codes implementation delay own the Code’s general history; what follows is the provident fund entry in that history, entry by entry.

The first movement was small and was driven by the pension litigation. By notification S.O. 2060(E) dated 3 May 2023, the Central Government brought into force a defined set of the Code’s provisions in so far as they related to the Employees’ Pension Scheme, 1995: sub-section (3) of Section 15, clauses of Section 16 concerning contributions to the Pension Fund, Section 143 in support of them, sub-section (1) of Section 164 to repeal the corresponding provisions of the 1952 Act specified in the third item of the repeal list, and clause (b) of sub-section (2) of Section 164 in so far as it related to the Pension Scheme, the clause that saves a scheme of a repealed enactment for the transition. The effect was precise: the provisions of the 1952 Act corresponding to the pension provisions of the Code stood repealed from 3 May 2023, the Pension Scheme itself was saved and continued, and the rest of the 1952 Act, its provident fund and insurance limbs included, was untouched.

The second movement was the general commencement. On 21 November 2025 the four labour codes were brought into force by notifications of that date, and the commencement notification for the Code on Social Security carried Section 164 into force in a manner that repealed eight of the nine listed enactments. The Employees’ State Insurance Act, the Payment of Gratuity Act and the remaining six fell. The 1952 Act did not. The notification’s entries for Sections 15, 16 and 164 carried exceptions keyed to the serial items of S.O. 2060(E), preserving the position created in May 2023 rather than superseding it, and no notification brought the repeal entry into force against the 1952 Act as a whole. A corrigendum issued on 19 December 2025 then corrected the commencement notification in terms, confirming that the provisions notified in 2023 stood excluded from the November repeal entries, so that the partial, pension-only repeal of 2023 and the survival of the remainder were placed beyond a drafting doubt. The clause of Section 164 that would enable the full transition of the provident fund schemes, the repeal of the Act itself, has not been notified for this Act. The dormancy is the more striking because the Code’s own provident fund chapter is a careful copy of the Act it would replace: Chapter III of the Code rebuilds the scheme-making power in Section 15, the contribution and fund provisions in Section 16, and the administration through the Central Board, in terms that track Sections 5, 6, 6A and 6C of the 1952 Act closely enough that the 2023 pension commencement could proceed by corresponding provisions. The replacement statute is drafted, enacted and in force; the machinery it would displace continues because the final notification, the one that would repeal the 1952 Act and move its schemes bodily into the Code, is the one notification the Government has chosen to withhold for this enactment.

The contrast with the eighth enactment on the list that did fall is instructive. The Employees’ State Insurance Act, 1948 was repealed at the general commencement, and its corporation, its contribution and its benefit structure passed into the Code’s corresponding chapter, a transfer made possible because an insurance scheme’s members hold entitlements to benefits rather than individual balances that must be preserved account by account. No equivalent transfer of the provident fund could be a mere change of letterhead at the statutory level, since the accounts, the first charge, the attachment protection and the determination machinery all hang on the 1952 Act’s sections, and each would have had to be re-enacted in the Code with savings of litigation-grade precision. The Government’s choice to move the pension provisions in 2023, when a judgment required it, and the schemes in 2026, when they were ready, while leaving the Act standing, reads in this light as a sequencing decision rather than an oversight.

What happened instead was a transition at the level of the schemes. Section 164(2) of the Code provides the bridge for schemes framed under a repealed enactment: they continue in force, to the extent they are not inconsistent with the Code, for a period of one year from the commencement of the Code or until replaced. Within that bridge period the Central Government exercised the scheme-making power in Section 15 of the Code, and by G.S.R. 525(E) dated 29 June 2026 framed the Employees’ Provident Funds Scheme, 2026 under Section 15(1)(a), superseding the Provident Fund Scheme of 1952 except as respects things done or omitted to be done under it, with companion schemes for pension and insurance framed under the same section. The 2026 scheme carried the substance across: the contribution structure of twelve per cent, ten per cent for specified establishments, and the wage ceiling of Rs 15,000 were retained, membership was continued for existing members, and accumulations and past actions were preserved.

The position that results can be stated in three sentences, and each has been earned by the sequence above. The 1952 Act is in force: its coverage, contribution, determination, recovery and priority provisions remain the operative law of provident fund compliance, and contributions continued to be computed under its definitions through the Code’s commencement and after it. The pension limb of the Act was repealed in May 2023 in the narrow corresponding-provisions sense, with the Pension Scheme saved, and pension has since been governed within the Code’s framework under the saved scheme and then its 2026 successor. The three schemes of 1952, 1995 and 1976 governed in their original form until the schemes of 2026 replaced them in form while continuing them in substance. A summary that says the EPF Act was repealed on 21 November 2025 is wrong on the notification record; a summary that says nothing changed is wrong on the pension entry of 2023 and the schemes of 2026. The truth is the sequence, and the sequence is the finding.

The Act among its neighbours: what the EPF Act 1952 is not

The provident fund is one instrument in a family of Indian social security devices, and its boundaries are clearest against its neighbours. The Employees’ State Insurance Act, 1948 is an insurance model: contributions from employer and employee buy defined medical and cash benefits administered by a corporation, and the member accumulates nothing; the entitlement is the benefit in the contingency, not a balance. The Payment of Gratuity Act, 1972 imposes no contribution at all: gratuity is a terminal payment computed on wages and service, borne by the employer alone, and its risk sits on the employer’s balance sheet until the day it falls due, which is why the insolvency protection for gratuity had to be written into the insolvency statute expressly. The coal mines fund of 1948 survives as a sectoral parallel, administered apart, for an industry whose provident fund predates the general Act. Against these, the provident fund’s identity is the individual account: a balance that is the member’s property in substance, portable across employers, inheritable through nomination, and fenced by Section 10 from the member’s own creditors. The pension scheme grafted onto the same contribution is the exception inside the design, a pooled promise riding on a savings statute, and the friction between the two logics, account and pool, is the source of nearly every hard question the Act has generated in the last decade, from the wage ceiling to higher pension to the design of the 2026 successor schemes.

One further contrast belongs in the map, with the market-linked pension system administered under the pension regulator’s statute. The National Pension System also builds an individual account, but it invests the account at market prices in instruments the subscriber chooses within regulated limits, carries no employer compulsion for the private sector, promises no rate and pays no defined benefit; the account’s value at retirement is whatever the accumulation buys. The provident fund account is the opposite temperament: a pooled investment, a declared rate smoothed across years, compulsion on both sides of the employment relation, and a pension arm carrying a formula. The two systems are frequently presented as rivals for the same savings, and at the level of a member’s monthly deduction they are; at the level of design they answer different fears, the one fearing the dead hand of a pooled fund and the other fearing the market at the moment of retirement, and the 1952 Act’s longevity owes something to the second fear being the one that organised labour has historically felt more strongly.

That friction also explains the Act’s political resilience. A pure pension statute can be redesigned by changing a formula; a savings statute holding the named balances of a very large share of the organised workforce cannot be disturbed without every member seeing the disturbance in a personal account. Governments have therefore amended the Act at the edges for seventy years and replaced its schemes rather than its shell, and the Code’s own drafters, for all their consolidating ambition, rebuilt the same three-fund structure inside Chapter III of the Code on Social Security. The 1952 Act’s survival of its own repeal provision is, in that light, less an accident of notification than the predictable fate of a statute whose repeal would have to be explained to every holder of an account.

Reading the payslip: the member’s audit of the EPF Act 1952 in practice

The analysis in this article reduces, for a member, to a small set of checks against the member’s own record, and they are worth stating as a practice. The first check is coverage and date: the account should open from the month the establishment stood covered and the member joined it, and a late opening is recoverable, since the employer’s duty runs from coverage and not from the day the organisation discovered the establishment. The second check is the wage base: the contribution should be twelve per cent of the Section 6 base, basic wages with dearness allowance and retaining allowance, and a base that is a small fraction of the total wage in a pay packet where the allowances are universal is the Bridge and Roof defect in payslip form. The third check is the split: the employer’s share should reach the account reduced by the pension diversion and by nothing else, and a payslip or passbook that shows the whole of the employer’s twelve per cent in the account is as much an error as one that shows less, since the missing diversion means the pension record is not being built. The fourth check is continuity: on a change of employer the universal account number should gather the new account under the old number, and the pension service should be carried by transfer or scheme certificate, because the pension formula pays for aggregated service and a broken record is paid for in years. The fifth check is the nomination: it should name the family as the scheme requires and should have been remade on marriage and on every change in the family the scheme treats as material. None of the five requires legal assistance to perform; each requires only the member’s own documents read against the statute, which is the use for which a pillar account of the statute exists.

The employer’s mirror image of the audit is a calendar rather than a set of documents. Coverage must be identified in the month the threshold is crossed and not in the month the organisation calls; the wage base must be settled once, on the Section 2(b) test as the courts have applied it, and applied uniformly rather than negotiated employee by employee; the remittance must leave by the fifteenth day after the month closes; the returns must reconcile to the payroll that an inspector will later read; and every transfer of the establishment, every contractor engagement and every restructuring of the pay packet must be tested against the section it engages before it is executed, since the Act prices each of those events and prices them retrospectively. An establishment run on that calendar rarely meets the determination machinery, and one run without it meets the machinery on the machinery’s terms.

The outer boundary: the workers the EPF Act 1952 does not reach

A pillar account must mark the boundary as plainly as the interior. The Act governs through establishments, and its universe is the organised sector: the worker with no establishment, the self-employed worker, the agricultural labourer on a farm that is not a covered establishment, and the domestic worker in a private household stand outside it, not by exemption but by the structure of Section 1(3), which fastens the obligation on an employer in an establishment of the stated size. The Code on Social Security is the first Indian statute to speak to the space outside, by building chapters for unorganised workers and for gig and platform workers alongside the provident fund chapter, funded on different principles and administered through boards rather than through employer deduction. The provident fund model was not extended to that space, and the design reason is visible in this article’s machinery: a contribution statute needs a payroll, an identifiable employer and a wage record, and where those are absent a different instrument, a welfare fund financed by cess or by aggregator contribution, must do the work. The boundary is therefore not an oversight in the 1952 design but the limit of its chosen mechanism, and the continuing exclusion of the unorganised majority from the account-based system is the largest single fact about Indian retirement provision, against which the Act’s internal controversies, the ceiling, the split and the higher pension option, are arguments among the covered.

What the survival of the EPF Act 1952 settles, and what it leaves open

The namable claim of this article can be put in its final form. The provident fund statute is an enactment listed for repeal in a Code that is itself in force, whose repeal was notified for every other listed enactment and not for it, so that the reader who consults the Code will conclude that the statute is gone and the reader who consults the notifications will find that it is not. This is a distinct category in the series’ catalogue of legislative afterlives, different from the statute never commenced and from the provision struck down but still printed: here the repealing instrument is fully alive, and only the repeal is dormant, for one named statute, by deliberate choice.

The choice is defensible, and its cost is visible. Continuity was preserved so that contributions, accumulations and pension service records were never disturbed by a change of statutory home; no member’s balance moved, no employer’s payroll arithmetic altered, and the transition, when it came in June 2026, arrived as a change of scheme letterhead rather than of substance. The cost is ascertainability. The governing position cannot be recovered from any single document: it requires the Code’s text, a notification of May 2023 issued for a pension judgment, a commencement notification of November 2025 read with its exceptions, a corrigendum of December 2025, and the scheme notifications of June 2026, and a payroll team that reads fewer than all of them will get the answer wrong in one direction or the other. The remedy is not a further statute but a table: a single authoritative commencement statement, maintained by the notifying ministry, that records for each listed enactment which repeal entries have been brought into force, on what dates, with what exceptions and savings, so that the question “is this Act in force” has one official answer instead of five documents in sequence. The absence of such a table is the administrative lesson of the episode, and it generalises: a codification that commences by exception and corrigendum multiplies the research burden on every regulated person, and the burden falls hardest on the smallest establishments, which are the ones least able to carry it and the ones the provident fund statute most exists to discipline.

Two questions remain open on the merits, and both can be answered without a statute. The first is the wage ceiling: a figure fixed in the prices of 2014 has been carried unchanged into a labour market whose organised-sector entry wages have moved well past it, so the band of wages on which the middle of the workforce contributes at the statutory minimum widens every year by drift rather than by decision, and the pension that the ceiling prices buys less for each succeeding cohort. An indexed ceiling would not settle the underlying argument between wider coverage and lighter compulsion, but it would force the argument to be held on purpose, at intervals, instead of being settled silently by inflation. The second is the pension promise’s funding: the diversion, the government contribution and the formula were fixed in different decades against different wage levels, and until a valuation-driven settlement ties the three together, the scheme’s generosity will continue to be decided in litigation about options rather than in policy about adequacy. Neither question is answered by the Act’s survival; both are made more visible by it, because a statute that has outlived its own repeal can no longer plead that its numbers were provisional.

For the member, the usable residue of the whole inquiry is simple. The Act applies, the contribution is twelve per cent of the statutory wage base from each side with the split this article has traced, the pension turns on the ceiling and on the option where it survives, and the fund’s protections, against the employer’s creditors, the member’s creditors and the employer’s insolvency, are exactly where Sections 10 and 11 left them. This article states the law for information; a member’s own position turns on the member’s service record, the scheme text and the facts of the establishment, on which advice should be taken.

Frequently Asked Questions

Q: Which establishments does the EPF Act 1952 apply to?

The Act applies under Section 1(3) to every establishment that is a factory engaged in an industry specified in Schedule I and employs twenty or more persons, and to any other establishment of a class notified by the Central Government employing twenty or more persons. Coverage once attached continues under Section 1(5) even if employment later falls below twenty, and an establishment below the threshold may be covered voluntarily under Section 1(4) by agreement with the majority of its employees. Section 16 keeps out specified cooperative societies working without power, and establishments whose employees receive equivalent provident fund or pension benefits under their own governing rules.

Q: What are the three schemes under the EPF Act?

The Act carries three schemes. The Employees’ Provident Fund Scheme, 1952, framed under Section 5, runs the individual savings account. The Employees’ Pension Scheme, 1995, framed under Section 6A, which was inserted by Act 25 of 1996 with effect from 16 November 1995, pools part of the employer’s contribution and a Central Government contribution into a Pension Fund that pays monthly pension. The Employees’ Deposit-linked Insurance Scheme, 1976, framed under Section 6C, inserted by Act 99 of 1976, pays an assurance benefit to the family of a member who dies in service, funded by the employer alone.

Q: How much is contributed and how is it split?

The contribution under Section 6 is twelve per cent of basic wages, dearness allowance including the cash value of any food concession, and retaining allowance, payable by the employee and matched by the employer; for notified classes the rate is ten per cent. Of the employer’s twelve per cent, eight and one-third per cent, capped at the wage ceiling, is diverted to the Pension Fund and the balance of three and two-third per cent reaches the provident fund account. The employee’s whole twelve per cent goes to the provident fund. The Central Government contributes 1.16 per cent of wages to the Pension Fund, and the employer separately pays one-half per cent for insurance and one-half per cent as administrative charges.

Q: What is the wage ceiling and why does it matter?

The wage ceiling is Rs 15,000 a month, in force from 1 September 2014. It is a figure of the schemes and not of the Act: the notifications of 22 August 2014, G.S.R. 608(E) for the Provident Fund Scheme and G.S.R. 609(E) for the Pension Scheme, substituted it for the earlier ceiling of Rs 6,500 that had stood from 1 June 2001. The ceiling matters three times over. It decides which joining employees must be covered, it caps the employer’s pension diversion at Rs 1,250 a month, and it fixes the pensionable salary from which the monthly pension is computed, unless a joint option supports contribution and pension on a higher salary.

Q: What did the Supreme Court decide about higher pension?

In Employees Provident Fund Organisation v. Sunil Kumar B., decided on 4 November 2022, the Supreme Court upheld the 2014 amendment of the Pension Scheme, including the wage ceiling, the fresh option for members contributing above it and the sixty-month average for pensionable salary. The Court held the extra contribution of 1.16 per cent of salary above the ceiling, exacted from members, to be beyond the parent Act, and suspended that invalidity for six months to allow the scheme to be adjusted. Members in service on 1 September 2014 who had missed the option received a further four months to exercise the joint option, and the facility was extended to members of exempted establishments.

Q: How are provident fund dues recovered?

Section 8 permits recovery of amounts due from an employer as arrears of land revenue. Sections 8B to 8G build a certificate machinery: the authorised officer issues a certificate to a Recovery Officer, who proceeds by attachment and sale of movable or immovable property, by arrest and detention of the defaulter, or by appointing a receiver, with Section 8G applying the attachment provisions of the Income-tax Act, 1961. Section 8F permits recovery from a person who holds money for the defaulter, and Section 8A lets a principal employer recover from a contractor the contributions paid for the contractor’s workers. The merits of the assessment cannot be reopened in the recovery forum.

Q: Is the EPF Act still in force after the labour codes?

Yes. Section 164(1) of the Code on Social Security, 2020 lists the Act among nine enactments for repeal, and the general commencement of 21 November 2025 repealed eight of the nine; the repeal entry was not brought into force against the 1952 Act as a whole. A narrower slice moved earlier: by S.O. 2060(E) of 3 May 2023, Code provisions were commenced in so far as they related to the Pension Scheme, the corresponding provisions of the 1952 Act stood repealed, and the Pension Scheme itself was saved. The schemes were replaced in form by the Employees’ Provident Fund Scheme, 2026, framed under the Code, with the twelve per cent rates and the Rs 15,000 ceiling retained.

Q: What protection does provident fund have in insolvency?

Section 11(1) gives provident fund dues priority over other debts where an employer is adjudicated insolvent or a company is wound up. Section 11(2), inserted by Act 40 of 1973, goes further and deems the amount due a first charge on the assets of the establishment, with priority over all other debts notwithstanding anything in any other law, a charge the Supreme Court applied ahead of secured creditors in the Esskay Pharmaceuticals case of 2011. In a liquidation under the Insolvency and Bankruptcy Code, 2016, Section 36(4)(a)(iii) keeps sums due to a workman or employee from the provident fund, pension fund and gratuity fund out of the liquidation estate altogether.

Q: Who is an excluded employee under the EPF Act?

An excluded employee is a creature of the Provident Fund Scheme, defined in paragraph 2(f). The term covers an employee whose pay at the time of joining exceeds the wage ceiling of Rs 15,000 a month and who has never been a member of the Fund, together with specified categories such as certain apprentices and a member who withdrew the full amount on retirement from a covered establishment after the age the scheme fixes. The definition works at the point of entry only. An employee who became a member while within the ceiling does not become excluded when pay later rises above it; membership, once attached, continues, with the contribution and the pension diversion capped at the ceiling figures unless a joint election carries them higher. International workers are not excluded employees.

Q: Can an establishment come under the EPF Act voluntarily?

Yes. Section 1(4) provides that where an agreement has been entered into between an employer and the majority of the employees, the Central Provident Fund Commissioner may, by notification in the Official Gazette, apply the Act to that establishment from the date of the agreement or a specified earlier date. Voluntary coverage is how an establishment below the twenty employee threshold, or outside the notified classes, comes within the Act without waiting for compulsion. Its consequences are the same as compulsory coverage: the contributions, the determination machinery and the offences all apply, and by Section 1(5) the coverage, once attached by either route, continues even though the number of employees later falls below twenty.

Q: What counts as basic wages for provident fund contributions?

Section 2(b) defines basic wages as all emoluments earned while on duty or on leave or holidays with wages, payable in cash under the contract of employment, excluding the cash value of any food concession, dearness allowance, house rent allowance, overtime allowance, bonus, commission and similar allowances, and presents. Section 6 then adds dearness allowance and retaining allowance back into the base on which the twelve per cent is computed. In Bridge and Roof Co. v. Union of India the Supreme Court held that an allowance paid universally and ordinarily to all employees forms part of basic wages, and in the Vivekananda Vidyamandir judgment of 28 February 2019 the Court held that special allowances paid universally across the establishment, without linkage to extra output or incentive, are basic wages on the same test.

Q: What interest and damages does an employer pay for delayed provident fund deposits?

Interest is charged under Section 7Q: the employer is liable to pay simple interest at twelve per cent per annum, or at the higher rate the Provident Fund Scheme specifies, for the period of the delay, and the liability is automatic in character, running from the due date to the date of deposit. Damages are levied separately under Section 14B, after giving the employer a reasonable opportunity of being heard, in an amount not exceeding the arrears. The Provident Fund Scheme grades the damages by the period of default, at five per cent per annum for delays under two months rising through ten and fifteen per cent to twenty-five per cent beyond six months, and the Central Board may reduce or waive damages for a sick industrial company under a sanctioned rehabilitation scheme.

Q: How is the monthly pension under the Employees’ Pension Scheme calculated?

The member pension is computed by the formula pensionable salary multiplied by pensionable service, divided by seventy. Pensionable salary is the average monthly pay drawn during the sixty months preceding the exit from service, computed by reference to the wage ceiling of Rs 15,000 unless a joint option supports a higher figure. Pensionable service is the period for which pension contributions were received, aggregated across employers where the service is carried by transfer or scheme certificate. A member needs ten years of eligible service for a monthly pension; shorter service draws a withdrawal benefit instead. Service of twenty years or more earns a weightage of two years in the formula, and a minimum pension of Rs 1,000 a month applies.

Q: Who administers the provident fund, and what does the Central Board of Trustees do?

Section 5A constitutes the Central Board of Trustees to administer the Fund. The Board is tripartite: it is chaired by the Union minister responsible for labour, with a vice-chairman, officials of the Central and State Governments, and representatives of employers and of employees in the numbers the scheme of the section lays down. Under Section 5C the Board is a body corporate with perpetual succession. Its functions run from administering the Fund and the three schemes to framing the subordinate arrangements for investment, accounts and claims, appointing officers, and issuing directions to the organisation. The executive machinery is the Employees’ Provident Fund Organisation, headed by the Central Provident Fund Commissioner appointed under Section 5D, working through regional and sub-regional offices and the Recovery Officers and inspectors the Act names.

Q: What happens to a member’s account when the member changes employment?

The account does not lapse and the membership does not end on a change of employer. The balance stands to the member’s credit, and on joining a covered establishment the member applies for transfer of the earlier accumulation to the new account; the universal account number links the member’s accounts across employments so that the record can be followed as one. For pension purposes the service under the earlier employer counts if it is carried by transfer or preserved through a scheme certificate, and the aggregation is what the pension formula rewards. A member who instead withdraws the provident fund on a short break in service settles the savings account, and must then attend to the pension record separately, since withdrawal of the fund and preservation of pensionable service are distinct choices with distinct consequences.

Q: What is the minimum pension payable under the Employees’ Pension Scheme?

The minimum member pension under the Pension Scheme is Rs 1,000 a month, fixed by the notification package of August 2014 with effect from 1 September 2014 and supported by budgetary provision from the Central Government. The floor applies to the member pension computed by the formula, so a member whose pensionable salary and service would produce a smaller amount is raised to the floor. The family pensions payable after a member’s death are expressed in the scheme as fractions of the member pension, widow or widower pension, children pension and orphan pension among them, and the operation of the floor among those derived pensions follows the scheme’s own text. The Rs 1,000 figure has stood unaltered since 2014, which is a standing item in the debate over the adequacy of the scheme.

Q: Can a member take an advance from the provident fund during service?

Yes. The Provident Fund Scheme permits advances and withdrawals from the member’s own accumulation during service for stated purposes, each carrying its own condition of service, ceiling and proof. The recognised purposes include marriage in the member’s family, the education of children, medical treatment of the member or family, the purchase or construction of a house or the repayment of a housing loan, and specified events of hardship such as closure of the establishment. Most of these advances are non-refundable withdrawals rather than loans, so the sum taken reduces the balance that earns interest and the lump sum payable at retirement, while leaving the pension record untouched. The claim is made through the employer or the organisation’s member interface, and the permanence of the reduction is the point a member should weigh before applying.

Q: What did the 2014 amendment change in the Employees’ Pension Scheme?

The amendment was made by notification G.S.R. 609(E) of 22 August 2014, effective 1 September 2014. It raised the wage ceiling from Rs 6,500 to Rs 15,000 a month, changed pensionable salary from a twelve-month average to an average of the sixty months before exit, required members contributing on salary above the ceiling to make a fresh option within six months to continue doing so, and required an additional contribution of 1.16 per cent of the salary above the ceiling from members who took that option. The same package fixed the minimum pension at Rs 1,000 a month. The extra 1.16 per cent member contribution was the element the Supreme Court struck down in 2022 as beyond the parent Act.

Q: How does the one year transition under the Code on Social Security work for the provident fund schemes?

Section 164(2) of the Code carries the savings. A scheme framed under a repealed enactment continues in force for one year from the commencement of the Code, to the extent it is not inconsistent with the Code, or until it is replaced by a scheme framed under the Code, whichever is earlier. For the provident fund, the replacement arrived within the window: the Employees’ Provident Fund Scheme, 2026 was framed under Section 15(1)(a) of the Code and notified as G.S.R. 525(E) on 29 June 2026, superseding the 1952 Scheme except as respects things done or omitted under it, with companion schemes for pension and insurance framed under the clauses that follow it. Past accumulations, service records and claims in motion were therefore carried across the change of letterhead without a fresh enrolment.

Q: What benefit does the Employees’ Deposit Linked Insurance Scheme pay on a member’s death?

The Insurance Scheme pays an assurance benefit to the nominee or family of a member who dies while in service in a covered establishment. The benefit is not a fixed sum: the scheme computes it from the member’s wages and the average balance in the member’s provident fund account over the period the scheme specifies, within a minimum and a maximum that the scheme fixes and that have been raised by notification over the scheme’s life. The cost is borne by the employer alone, at one-half per cent of wages under the scheme, within the one per cent ceiling of Section 6C; the member contributes nothing. Payment is made with the provident fund and pension death claims, to the payee the scheme’s nomination and family rules identify.