On 1 February 2017 the Finance Minister of India, Arun Jaitley, rose in the Lok Sabha to present the Union Budget and used the occasion to announce a new instrument of party funding called the electoral bond. A purchaser would buy the instrument from a bank with money drawn from a taxed and recorded account, hand the physical instrument to a political party, and see the amount credited to that party through the banking system, yet no voter, no rival party and no election authority outside the issuing bank would ever learn the purchaser’s name. Seven years later, on 15 February 2024, a five judge Constitution Bench of the Supreme Court, presided over by Chief Justice D. Y. Chandrachud, struck the entire arrangement down as a violation of the voter’s right to information under Article 19(1)(a) of the Constitution. Between the announcement and the annulment, bonds worth 16,518 crore rupees were sold by the State Bank of India. This article defends one claim about that trajectory. The electoral bond was not best understood as a reform proposal that happened to fail, nor as a corruption story that happened to end in court. It was an architecture, a designed system of four interlocking legal components whose joint product was anonymity, and the events of 2018 to 2024 tested that architecture component by component until the Constitution Bench dismantled the whole of it and the disclosure of March 2024 published the record the architecture had been built to withhold.

The Legislative Moment: How Electoral Bonds Entered Law in 2017

The budget speech of 1 February 2017 deserves to be read as the founding document of the scheme, because almost everything that followed was an execution of sentences spoken that afternoon. Jaitley told the Lok Sabha that even after seventy years of independence the country had not evolved a transparent method of funding political parties, that parties received most of their funds in cash from anonymous sources, and that donors hesitated to give by cheque because the disclosure of their identity invited adverse consequences from rival parties. His remedy had four limbs. The cash donation a party could accept anonymously would fall from 20,000 rupees to 2,000 rupees. Donations above that figure would have to arrive by cheque or digital transfer. An amendment to the Reserve Bank of India Act 1934 would authorise the government to notify a scheme under which a donor could purchase bonds from scheduled banks and give them to a party of the donor’s choice. And the party would redeem the instrument into a single notified bank account, so that every rupee of bond money would travel through the banking system while the name attached to it travelled nowhere outside the issuing branch.

What did the budget speech promise donors in return for giving through banks?

Jaitley promised donors that if they routed political giving through the banking system by buying electoral bonds, their identity would be shielded from the recipient party’s rivals, from the public and from disclosure, so that the fear of adverse consequences would no longer push honest money back into cash.

The promise is worth stating precisely because the constitutional case of 2024 turned on it. The budget speech did not present anonymity as an incidental side effect of a banking instrument. It presented anonymity as the inducement, the consideration offered to the donor in exchange for surrendering cash. Cash was untraceable and therefore safe for the giver but useless for public accountability. A cheque was traceable and therefore accountable but unsafe, in the government’s account, for the giver. The bond was engineered to break that trade off. It would be traceable as money, because no bond could be bought except through a know your customer compliant account, and untraceable as allegiance, because the law would forbid every institution except the issuing bank from connecting the money to a name, and would forbid the bank from publishing the connection. The scheme thus separated two kinds of knowledge that the older law of political finance had kept together. The state would know that money had moved. Nobody outside one bank’s ledger would know whose money it was or which party’s politics it rewarded.

Parliament enacted the framework within weeks, and the route it took through Parliament became part of the controversy that reached the Supreme Court. The Finance Bill 2017 was certified as a Money Bill, which confined the Rajya Sabha to recommendations the Lok Sabha could disregard, and it carried amendments to four statutes that had nothing to do with taxation in the ordinary sense. Section 31 of the Reserve Bank of India Act was amended to let the central government authorise a bank to issue electoral bonds. Section 29C of the Representation of the People Act 1951, the provision that obliges every party to report contributions above 20,000 rupees to the Election Commission, was amended so that contributions received through bonds fell outside the report altogether. Section 13A of the Income Tax Act 1961, which grants parties exemption on voluntary contributions provided they keep records of donors above the reporting threshold, was amended so that no record of a bond donor’s name or address need be maintained. Section 182 of the Companies Act 2013 was amended to delete the ceiling on corporate political contributions and to delete the duty to name recipient parties in the profit and loss account. Each amendment is examined in its own section below, because each is one component of the architecture. The point to register at the legislative moment is that anonymity was not located in any single clause. It was distributed across four codes, so that removing any one lock would still leave the others fastened.

Two companion changes completed the design and are often missed in summaries that treat the bond as a stand alone instrument. The Finance Act 2016 had already amended the Foreign Contribution (Regulation) Act 2010, retrospectively to 2010, to redefine a foreign source so that an Indian company with foreign shareholding within the sectoral limits of foreign direct investment policy was no longer a foreign source for donation purposes, and the Finance Act 2018 extended that redefinition retrospectively to 1976, the year the original Act commenced. The effect was that the foreign owned yet Indian incorporated donor, previously barred, could lawfully fund parties, and could do so through the same anonymous channel after 2018. The second change was the cash ceiling itself. By cutting the anonymous cash limit to 2,000 rupees in section 13A while opening an unlimited anonymous banking channel beside it, the Finance Act 2017 did not merely offer donors a cleaner route. It repriced the dirtier one. A donor who wished to give a crore of rupees without a name attached could no longer split the gift into cash tranches below the old threshold without absurd labour. The bond was the only remaining instrument of unnamed giving at scale, and it was deliberately built that way.

The notification that converted the statute book into a working market came ten months later. The Electoral Bond Scheme, 2018 was notified in the Gazette on 2 January 2018, and the first bonds were sold in March 2018. The notification, not the budget speech, fixed the features that buyers and parties actually handled: the State Bank of India as sole issuer, five denominations running from 1,000 rupees to 1 crore rupees, a validity of fifteen calendar days, sale windows of ten days in each quarter, and eligibility confined to parties registered under section 29A of the Representation of the People Act that had secured at least one percent of the votes at the preceding general election. The scheme operated for thirty sale phases across six years. It was challenged in court before the first phase had closed, defended by the government through four rounds of hearings, preserved by interim orders that traded secrecy against sealed cover deposits with the Election Commission, and finally struck down in a unanimous judgment that directed the bank to publish what it held. To follow that arc with the precision it deserves, the article first reconstructs the longer history of Indian political finance that the scheme claimed to cure, then takes the architecture apart piece by piece.

A Long Prehistory: Party Money Before the Bond

The claim that India had failed to evolve a transparent method of funding parties was, by 2017, nearly seventy years old, and each earlier failure left a deposit in the law that the bond scheme either used or removed. The Representation of the People Act 1951 regulated candidate expenditure from the first general election but left party finance largely untouched, on the theory that parties would raise modest sums from members. Corporate money entered early and lawfully. The Companies Act 1956, through section 293A, permitted a company to contribute up to five percent of its average net profits to political parties, with disclosure in the accounts. The arrangement lasted until 1969, when the Companies (Amendment) Act of that year banned corporate contributions entirely, in a Parliament persuaded that company money was purchasing policy. The ban did not end corporate funding. It drove funding into cash and into the parallel economy that the Wanchoo Committee on direct taxes, reporting in 1971, identified as a principal consumer of unaccounted money, and the connection between election finance and black money became a fixed finding of official inquiries for the next four decades.

The ban was lifted in 1985, when the government of Rajiv Gandhi amended the Companies Act to restore corporate contributions at the five percent ceiling with board authorisation and disclosure. The restoration admitted what the ban had taught: a democracy with expensive elections and no lawful corporate channel will be financed by an unlawful one. Through the 1990s the Election Commission under T. N. Seshan enforced candidate expenditure ceilings with unprecedented rigour, a period examined in the account of the Seshan era reforms at the Commission, but party income sat outside the expenditure ceiling’s reach, and the gap between what campaigns visibly cost and what parties declared they spent became the standing embarrassment of the system. The Supreme Court addressed income in Common Cause v Union of India in 1996, holding that parties were obliged to maintain and file accounts and that unexplained party expenditure could be attributed. The Indrajit Gupta Committee on State Funding of Elections, reporting in December 1998, recommended a partial state subvention in kind, free airtime and supplies rather than cash, and Parliament enacted the airtime element for recognised parties. The Law Commission’s 170th Report on Reform of the Electoral Laws, submitted in May 1999 under Justice B. P. Jeevan Reddy, proposed a regulatory statute for parties themselves, with audited accounts and disclosure as the price of recognition.

Disclosure law advanced furthest through litigation about candidates rather than donors. In Union of India v Association for Democratic Reforms, decided on 2 May 2002, the Supreme Court directed the Election Commission to require every candidate to declare criminal antecedents, assets, liabilities and educational qualifications, grounding the direction in the voter’s right to know under Article 19(1)(a). The People’s Union for Civil Liberties judgment of 13 March 2003 upheld the statutory version of that disclosure against a parliamentary attempt to dilute it. Those two judgments matter to the bond story for a doctrinal reason that the 2024 bench would make explicit. They established that information bearing on the voter’s choice is not a policy concession granted by Parliament but a constituent of the Article 19(1)(a) freedom itself. When the petitioners of 2017 argued that information about who funds a party stands on the same footing as information about a candidate’s record, they were not inventing a right. They were extending a line of authority the Court had drawn twice, and the government that opposed them had to explain why the line stopped at the party’s bank account.

The immediate prehistory of the bond is the data on unknown sources. The Association for Democratic Reforms, analysing the income tax returns and contribution reports that parties filed with the Election Commission, calculated that between the financial years 2004-05 and 2014-15 roughly sixty nine percent of the declared income of national parties came from sources the parties were not required to name, because each contribution fell below the 20,000 rupee reporting threshold of section 29C or arrived as cash coupons and donations in kind. Successive Law Commission work, including the 255th Report on Electoral Reforms of March 2015, chaired by Justice A. P. Shah, recommended lowering the disclosure threshold and tightening the law on party accounts, and the Election Commission’s own compendium of proposed reforms repeated the demand. The direction of expert travel for two decades was therefore unambiguous: lower thresholds, wider disclosure, tighter audit. The Finance Act 2017 moved on a different vector. It lowered the cash threshold, which the reform literature welcomed, and in the same statute built an instrument whose governing rule was that no threshold question could ever arise, because the contribution would not enter the disclosure system at all. That doubleness, one clause tightening the old regime and four clauses exempting the new instrument from it, is the legislative fact that the architecture reading keeps in view and that framings organised around reform or around scandal both tend to blur.

Component One: Donor Anonymity Through the State Bank of India

The first component of the architecture was the channel. The notification of 2 January 2018 designated the State Bank of India, the country’s largest public sector bank, as the only institution authorised to issue and encash electoral bonds, initially through four branches in Chennai, Kolkata, Mumbai and New Delhi and eventually through twenty nine authorised branches. A bond could be purchased only by a citizen of India or by an entity incorporated or established in India, only through a know your customer compliant account, and only by cheque or electronic transfer, never for cash. The instrument itself was a bearer instrument in the nature of a promissory note, interest free and payable to the bearer on demand. It carried no name. The purchaser’s name existed in exactly one place, the issuing branch’s records of the account from which payment had been made, and the scheme imposed no duty on the bank to reveal that record to any person, any party or the Election Commission.

Who was allowed to sell electoral bonds, and through how many branches?

Only the State Bank of India could issue or encash an electoral bond. The bank began through four authorised branches in Chennai, Kolkata, Mumbai and New Delhi in March 2018, and the network of authorised branches grew to twenty nine as the scheme’s sale phases continued.

Consider what this design did to the distribution of knowledge. In the older cheque regime, three parties to a donation knew the donor’s identity: the donor, the bank processing the cheque, and the recipient party, whose own accounts carried the name. The bond regime rearranged that triangle. The issuing bank knew. The recipient party, receiving a bearer instrument identical in form to every other instrument of its denomination, did not know from the instrument itself, and the law required the party to record nothing that would tell it. The Election Commission, which under section 29C received contribution reports naming every donor above 20,000 rupees, was statutorily excluded from receiving this class of contribution in any named form. The voter, whose Article 19(1)(a) right the 2002 and 2003 judgments had built, stood outside the arrangement entirely. Anonymity under the scheme was therefore not the absence of records. It was an allocation of records, carefully drawn so that the single institution holding the complete picture was a bank owned by the state whose government of the day was also the largest prospective beneficiary of donations, and so that no institution charged with conducting elections or informing voters held any picture at all.

The government defended this allocation as the scheme’s virtue, and the defence must be taken at full strength because the Constitution Bench took it seriously before rejecting it. The argument, made first in the budget speech and repeated by the Attorney General in the hearings of October and November 2023, ran as follows. Donors to opposition parties in a state could face retaliation from the government of that state, and donors to a ruling party could face retaliation when power changed hands. Fear of that retaliation is what sustains cash. A donor protected against exposure will give through banks, and banked money is at least money whose quantity is known, whose source account is taxed, and whose path can be reconstructed by a court order or an investigating agency with lawful authority. On this account the scheme did not abolish knowledge. It converted public knowledge into official knowledge, available to the state on demand, and traded the voter’s sight of donations for the taxman’s grip on them. The petitioners’ reply, which the lead judgment substantially accepted, was that the trade was illusory in one direction and real in the other. The state’s grip depended on the state choosing to grip, and the state in question was the incumbent government, the one actor with an institutional interest in not reconstructing the funding of itself and its friends. The voter’s loss, by contrast, was total and automatic, operating in every election held while the scheme lived.

The eligibility rule at the channel’s far end compounded the allocation of advantage. A party could encash bonds only if it was registered under section 29A and had secured at least one percent of the votes polled at the most recent general election to the Lok Sabha or to a state legislative assembly. The threshold sounded modest, and for national and established regional parties it was no obstacle at all. Its effect fell on the new and the small, the entrants for whom a single large donation might finance a first serious campaign, and who were told that the anonymous channel was open to their donors only after the party had already demonstrated electoral weight without it. Established parties, possessing the vote shares the rule demanded, faced no such gate. The one percent rule therefore worked as an incumbency filter inside the instrument itself, independent of any party’s conduct, and it aligned with the pattern the disclosed data later displayed, in which parties holding government office, at the centre and in the states, absorbed the overwhelming bulk of the money the channel carried. A design defended as neutral plumbing between donor and party had a gradient built into its pipes, and the gradient ran toward power.

A second design choice inside the first component sharpened that dispute, and it surfaced in the disclosed data of 2024. The bond was a bearer instrument, transferable by delivery. The Reserve Bank of India had warned the Finance Ministry in January 2017, in correspondence later obtained under the Right to Information Act and placed before the Supreme Court by the petitioners, that bearer instruments are notoriously opaque, that who finally contributes such an instrument to a party cannot be known even in principle, and that issuing them in sizeable quantities could set a bad precedent by encouraging money laundering. The bank’s central board committee proposed that if the scheme proceeded, the instruments should exist in electronic or dematerialised form, so that each transfer left a record. The government kept the paper. The consequence is visible in the architecture itself. Even the issuing bank’s ledger, the one complete record the scheme contemplated, recorded purchases rather than donations. Between purchase and encashment a bond could change hands, and the ledger would be silent about the hands. The anonymity the scheme sold was thus deeper than its own defence described, because the defence assumed a purchaser who was also the donor, and the bearer form dissolved that assumption.

Component Two: Receipt Without Attribution

The second component governed the receiving end. A party eligible under the scheme, registered under section 29A and having secured at least one percent of the votes polled at the most recent general election to the Lok Sabha or to a state legislative assembly, maintained a designated account with the issuing bank, and a bond deposited into that account within its fifteen day validity was credited at face value. So far the arrangement resembled an ordinary banking service with an eccentric instrument. The legal transformation sat in the reporting statutes. The proviso that the Finance Act 2017 inserted into section 29C of the Representation of the People Act exempted contributions received by way of electoral bond from the contribution report that parties must file with the Election Commission. The amendment to section 13A of the Income Tax Act relieved parties of the duty to keep a record of the name and address of any person contributing through a bond, a duty that otherwise attaches to every voluntary contribution above 20,000 rupees as a condition of the party’s tax exemption.

The combined effect deserves slow statement, because the aggregate figures that parties did publish under the scheme have misled many summaries into describing the scheme as one of partial disclosure. A party’s audited annual accounts, filed with the Election Commission, showed a total figure for donations received through bonds. The country could therefore learn, usually more than a year after the fact, that a party had received a stated sum through the channel. What no filing anywhere revealed was the composition of that sum. A total of six hundred crore rupees might represent six hundred donors of one crore or one donor of six hundred crore, an industrial house purchasing favourable policy or a multitude of small supporters, a single transaction in the week before a contract was awarded or an accumulation across six years. Every question a voter might ask about influence requires composition, and composition is precisely what component two removed from the record. The architecture did not darken the theatre. It left the lights on and removed the cast list.

The Election Commission identified this defect before a single bond had been sold, and identified it in the vocabulary of its own statute. In its letter of 26 May 2017 to the Ministry of Law and Justice, written within weeks of the Finance Act’s passage, the Commission stated that the amendments would have a serious impact on the transparency of political funding, called the section 29C exemption a retrograde step as far as transparency of donations is concerned, and asked that the proviso be withdrawn. The letter’s second objection was structural rather than rhetorical. Section 29B of the Representation of the People Act forbids parties to accept contributions from government companies and from foreign sources. If contributions arriving through bonds are reported nowhere, the Commission reasoned, it cannot be ascertained from a party’s contribution report whether the party has taken a donation that section 29B prohibits. The exemption therefore did not merely hide lawful information. It disabled the enforcement of a prohibition that remained on the statute book, by destroying the evidential trail on which enforcement depended. A reader weighing the framings this article adjudicates should hold that sentence of the Commission’s letter in mind, because it was written by the institution the Constitution charges with superintendence of elections, it was written before the scheme existed in operation, and events never gave the Commission cause to withdraw it. The Commission repeated the substance in its counter affidavit filed in the Supreme Court on 27 March 2019, in the middle of a general election campaign financed substantially through the instrument it was criticising.

The contrast with the older instruments of party finance measures what component two changed. Electoral trusts, the vehicle introduced by the Electoral Trusts Scheme of 31 January 2013, collected contributions from companies and persons and distributed them to parties, and the trusts were required to report their donors and their distributions to the Election Commission, so that the trust route pooled money without anonymising it. A donation by cheque was named in the donor’s accounts, the party’s records and, above the threshold, the section 29C report. Even the much criticised regime of sub threshold cash at least preserved a legal fiction of traceability, because a party claiming that six hundred crore rupees arrived in donations of 19,999 rupees asserted a fact that investigators could test against the party’s circumstances. The bond alone created a class of contribution that was large by design, banked by design, and unattributable by law. Component two is where the scheme’s critics located the constitutional injury, and the 2024 judgment agreed with their location. The lead opinion held that information about funding to a political party is essential for a voter to exercise the freedom to vote in an effective manner, and that the scheme and the amended provisions, to the extent that they anonymised contributions, violated Article 19(1)(a). Receipt without attribution was not a bookkeeping detail. In the Court’s analysis it was the point at which the architecture touched the franchise.

Component Three: Denominations, Fifteen Days and the Windows of Sale

The third component was the instrument’s physical and temporal shape, and it repays attention because shape determined use. The notification of 2 January 2018 fixed five denominations: 1,000 rupees, 10,000 rupees, one lakh rupees, ten lakh rupees and one crore rupees. A bond remained valid for fifteen calendar days from the date of issue, within which the payee party had to deposit it into its designated account, failing which the amount lapsed to a fund administered under the scheme’s terms. Bonds were sold in windows of ten days in January, April, July and October of each year, with an additional thirty days in a year in which a general election to the Lok Sabha was held, and the government retained the power to notify further windows, a power exercised in the run up to state elections on several occasions, including an additional window in 2022 and 2023 that the petitioners challenged as timed to assembly polls.

How long did a purchaser have to place an electoral bond with a party?

Fifteen calendar days ran from the date of issue. A bond not deposited by an eligible party into its designated account inside that period expired, and the issuing bank credited the lapsed amount to the fund the scheme maintained for such unclaimed sums.

Each feature of the shape pushed the instrument toward a particular kind of transaction. The one crore denomination, the largest, made the bond an instrument of wholesale finance. In the data the State Bank ultimately disclosed, purchases clustered overwhelmingly at the highest denomination, which is to say the channel was used as it was built to be used, by buyers moving sums that only companies and very wealthy persons command. The fifteen day validity prevented the bond from circulating as a store of value or a parallel currency, answering part of the Reserve Bank’s anxiety about bearer instruments, but it also imposed a choreography on giving. A purchaser had to know, within a fortnight, which party would receive the instrument, and the party had to bank it promptly. Donations under the scheme were therefore planned events, coordinated between buyer and recipient in advance, not spontaneous offerings. The quarterly windows reinforced the choreography. Giving was batched into seasons announced by the government, and the seasons could be extended by the same government when an election approached, which meant the incumbent executive controlled the calendar of the instrument through which it was itself principally financed. Opposition parties made this point in Parliament and in their petitions, and the disclosure data of March 2024 gave it an empirical edge, because purchase dates could then be set against the election calendar and against the dates of government decisions affecting the purchasers.

The shape also explains a feature of the litigation that puzzles readers who meet the case only through its final judgment. Because validity was fifteen days and windows were quarterly, the scheme’s operation was episodic, and each episode closed before a court could realistically examine it. Petitions filed in 2017 challenged the framework before any bond existed. Stay applications pressed in 2019 confronted a court with a general election already under way and a channel already loaded with over a thousand crore rupees of instruments. The architecture’s tempo, its short windows and shorter validity, functioned as a procedural shield, and the Supreme Court’s interim order of 12 April 2019, which declined to stay the scheme while ordering parties to deposit their bond receipts, donor details and bank account particulars with the Election Commission in sealed covers, preserved the evidence while permitting the market to continue. That order, little noticed at the time, is the hinge between the scheme’s life and its posthumous exposure, because the sealed cover deposits it compelled became part of the record the 2024 judgment ordered open.

Component Four: Section 182 and the End of the Corporate Ceiling

The fourth component was not in the notification at all. It sat in company law, and without it the first three components would have processed a far smaller volume of money. Section 182 of the Companies Act 2013, as enacted, permitted a company to contribute to a political party up to seven and a half percent of its average net profits during the three preceding financial years, required a resolution of the board of directors authorising the contribution, and required the company to disclose in its profit and loss account both the total amount contributed and the names of the parties to whom it was given. The Finance Act 2017 deleted the ceiling, deleted the requirement to name the recipient parties so that only an aggregate figure need appear, and left the board resolution as the principal internal formality. A company could thereafter lawfully contribute any sum its board approved, including a sum exceeding its profits, including a sum contributed by a company with no profits at all, and could report the outgo in its accounts as an unnamed total.

The Election Commission’s letter of 26 May 2017 foresaw the consequence with uncomfortable exactness. Removing the profit linked ceiling, the Commission warned, opens the possibility of shell companies being set up for the sole purpose of making donations to political parties, with no other business consequence. The warning identified the mechanism. A ceiling fixed as a fraction of profits tethers political giving to commercial substance. A company must have traded, and traded successfully across three years, before it can give at scale, and the size of its giving is bounded by the size of its success. Delete the tether and a letterbox company with paid up capital of a few lakh rupees can receive funds from an undisclosed principal and pass them to a party through the bond channel, leaving the bank’s ledger showing a lawful purchase by a registered entity. The disclosed data of 2024 contained exhibits for the Commission’s prediction. Analysts working through the State Bank’s lists found companies whose bond purchases exceeded their recorded profits across entire years, companies with negligible paid up capital purchasing bonds worth hundreds of crore rupees, and at least one logistics firm whose purchases in a single financial year ran to many multiples of its net profit for that year. Section 182 is treated at greater length in the answers at the end of this article, because the Constitution Bench’s treatment of it contains the judgment’s only finding of manifest arbitrariness, a doctrine with a demanding threshold. The bench held that the amendment to section 182 was manifestly arbitrary, reasoning that it treated companies and individual persons alike for the purpose of political contribution although the two stand differently in relation to the electoral process, that it permitted unlimited contributions by corporate entities whose owners and controllers are invisible behind the corporate form, and that the prospect of loss making companies contributing at scale confirmed that the amendment served purposes unconnected with the health of corporate governance it was supposed to inhabit.

The shareholder dimension of component four drew less public attention than the shell company dimension, and the judgment’s reasoning gives it a weight commentary has not always matched. A company’s political contribution is decided by its board and, behind the board, by the promoters or managers who control it, but the money contributed belongs to the company, which is to say to all its shareholders in proportion to their holdings, including shareholders who support the recipient party’s opponents or who bought their shares as a pension investment with no wish to fund any party at all. The 2013 Act’s twin requirements answered that problem in the only way company law can: the ceiling bounded the sum that could be diverted from commercial purposes, and the naming duty let a shareholder learn which parties the board had funded and sell, object or organise accordingly. Deleting both requirements in a single Finance Act clause severed the shareholder from the information and the limit at once, and did so for contributions routed through an instrument that hid the gift from the public as well. The lead judgment’s observation that the amendment placed corporate and individual contributors under one rule despite their different constitutional situations rests substantially on this asymmetry: an individual’s gift expresses the giver, while a company’s gift expresses its controllers at its owners’ expense.

The interaction between component four and the other three is the architecture’s engine. Component one supplied a channel that hid the purchaser. Component two supplied a receipt rule that hid the composition of party income. Component three supplied denominations that made the channel wholesale. Component four removed the ceiling that would have limited how much corporate money the wholesale channel could carry, and removed the naming rule that would have let shareholders and the public see where the money went. Remove any one component and the scheme leaks information or volume. Together they formed a closed system in which the largest political contributions in Indian history could be, and were, made in complete legality and complete obscurity, with every institution of the state performing its assigned role: the bank recording what it would not publish, the Commission receiving totals it could not decompose, the tax authority exempting income it could not attribute, and the companies regulator accepting accounts that named no recipient. Students of the period reach for the language of institutional failure. The architecture reading suggests the colder conclusion that the institutions performed as redesigned.

The Artifact: The Four-Component Anonymity Architecture Map

The analysis so far can be compressed into a single citable instrument, and compression matters because the debate about electoral bonds has suffered from summaries that describe the scheme as a bank product rather than as a system of legal parts. The instrument is called the Four-Component Anonymity Architecture Map. Its rows are the four components examined above. Its columns record the legal vehicle that carried each component, the function the component performed inside the system, and the fate of that component in the judgment of 15 February 2024. A reader who grasps the map grasps why the Constitution Bench could not save the scheme by striking one clause, because each row’s function presupposed the rows around it, and why the post March 2024 disclosure carried such evidentiary force, because the disclosure was simply the map’s first two rows run in reverse, the bank’s withheld record published at last.

Component	Legal vehicle	Function inside the system	Fate on 15 February 2024
Donor anonymity	Electoral Bond Scheme notification of 2 January 2018, issued under section 31 of the Reserve Bank of India Act as amended	Purchase confined to the State Bank of India, identity recorded by the bank alone, bearer instrument carrying no name	Scheme struck down as violative of Article 19(1)(a)
Receipt without attribution	Proviso to section 29C of the Representation of the People Act 1951 and the proviso in section 13A of the Income Tax Act 1961	Bond contributions excluded from the contribution report to the Election Commission and from donor record keeping	Amended provisions struck down to the extent they anonymised contributions
Denominations and validity	Scheme notification of 2 January 2018	Five denominations from 1,000 rupees to 1 crore rupees, fifteen day validity, quarterly sale windows	Fell with the scheme; issuance directed to stop immediately
Corporate ceiling removed	Section 182 of the Companies Act 2013 as amended by the Finance Act 2017	Profit linked ceiling of seven and a half percent deleted, duty to name recipient parties deleted	Amendment held manifestly arbitrary and unconstitutional

The map also disciplines the counterfactual questions that dominate commentary on the verdict. Could the Court have preserved bonds while ordering disclosure? The map shows that disclosure is not an accessory bolted onto the scheme but the negation of rows one and two, so that a disclosing bond scheme is a different instrument wearing the same name, essentially a cheque with extra steps. Could Parliament repair the scheme with a disclosure threshold, sparing small donors and naming large ones? That question, which the judgment leaves open and the conclusion of this article takes up, is a question about designing a new map rather than amending the old one, and it is best debated in those terms.

The Case for the Scheme, Stated in Its Own Terms

An adjudication that does not state the losing side at full strength persuades nobody worth persuading, and the government’s case for the bond was neither frivolous nor concealed. It rested on three propositions, each with a pedigree. The first was diagnostic. Political funding in India ran on cash because cash was the only reliably secret medium, and the secrecy was not incidental to the corruption but constitutive of it, since unaccounted money enters politics precisely where no ledger exists. Jaitley returned to the argument after the budget, in a January 2019 essay and in parliamentary exchanges, maintaining that the choice before the country was never between the bond and some Norwegian transparency but between the bond and the suitcase. The second proposition was protective. Donors face consequences for disclosed allegiance in a polity where governments control contracts, licences, inspections and credit, and where power alternates. The Attorney General’s written note of October 2023, submitted on behalf of the Union during the Constitution Bench hearings, pressed the point to its limit by contending that citizens have no right to know the source of funding of a political party at all, that the judgments on candidate disclosure concerned the candidate’s own antecedents and could not be extended to the party’s benefactors, and that a donor’s preference for an unnamed gift is itself an aspect of political privacy deserving constitutional room.

Why did the Attorney General argue that voters had no right to know who funds a party?

Attorney General R. Venkataramani’s note of October 2023 argued that the Article 19(1)(a) right to information recognised in the candidate disclosure cases covered a candidate’s antecedents and assets, not the identity of a party’s donors, whose political preference the note treated as private.

The third proposition was comparative and institutional. Other democracies, the government noted, struggle with the same triangle of money, secrecy and influence, and several protect donor identity below stated thresholds precisely to keep small and middling givers inside the lawful system. The Solicitor General, Tushar Mehta, appearing for the Union in the hearings of 31 October to 2 November 2023, added a separation of powers argument with genuine force: that the design of an electoral finance regime is economic and political policy of the kind courts ordinarily leave to legislatures, that the judiciary lacks the institutional equipment to calibrate thresholds and channels, and that a court which strikes down a finance scheme assumes responsibility for the vacuum it creates without the means to fill it. He also drew on the comparative embarrassment that the petitioners’ own preferred regimes had never been operated in India at Indian scale. Each of these arguments was answered, and the judgment’s reasoning is set out below by holding. But the architecture reading insists on recording them at strength for a further reason. The verdict did not hold that donor privacy is a fiction or that retaliation is imaginary. It held that the specific design before it failed the proportionality enquiry, because total anonymity for unlimited sums was not the least restrictive means of protecting a donor, and because the interest asserted on the voter’s side was not curiosity but the functioning of the franchise. A reader who skips the government’s case will misread the judgment as a moral verdict. It is better read as an engineering judgment about a machine that protected one real value by destroying another.

Context for the protective argument sits inside the post-2014 PMO architecture context, where the concentration of executive discretion over contracts and clearances is documented as a structural feature of the period rather than an allegation against any office holder. The same concentration is what made the retaliation fear credible to donors and what made the anonymity valuable to the incumbent, and that doubleness runs through the entire episode. Anonymity protected the opposition’s donors from the government and the government’s donors from the public in a single motion, and the two protections were not equally distributed, because the government of the day controlled the bank’s regulator, the tax authority and the investigative agencies whose forbearance the anonymity was also purchasing. The petitioners put the asymmetry plainly. The scheme did not create a level field of secrecy. It created a field on which one player, the ruling party at the centre, could both raise the most money through the channel and command the only institutions capable of piercing it, a point the bench returned to when it observed that the scheme’s design tilted toward the party in power because the governing party could access the bank’s information through its control of the state in ways denied to every voter.

The Objections Filed Before the First Bond Was Sold

The challenge to the scheme was not retrospective wisdom assembled after the disclosure of 2024. Its principal arguments were placed on official record in 2017, by constitutional functionaries writing before the scheme operated, and the antiquity of the objections matters to any honest adjudication between framings. The Reserve Bank wrote first. On 28 January 2017 the Finance Ministry sought the bank’s comments on amending section 31 of the Reserve Bank of India Act to permit the instruments, and the reply, sent on 30 January 2017 under the authority of Governor Urjit Patel, opposed the design on grounds that read in retrospect like a table of contents for the litigation: bearer instruments are transferable by delivery, so the final contributor cannot be known; such instruments have the potential to become currency; issuing them in sizeable quantities can undermine faith in banknotes; and the amendment would set a bad precedent by encouraging money laundering and eroding a core principle of central banking legislation. The bank’s internal committee urged that any such instrument be issued in electronic form only. The Finance Ministry’s revenue secretary dismissed the objection in a note recording that the bank appeared not to have understood the mechanism, whose entire point was prepaid instruments purchased from tax paid money with the donor’s identity kept secret. The episode, reconstructed by journalists from files disclosed under the Right to Information Act in November 2019, fixed the documentary record: the state’s own monetary authority had named the bearer form as the defect, and the defect had been retained with knowledge.

What did the Election Commission tell the Law Ministry on 26 May 2017?

The Commission wrote that the Finance Act amendments would have a serious impact on transparency in political funding, called the exclusion of bond contributions from the section 29C report a retrograde step, warned that prohibited donations could no longer be detected, and asked for the proviso to be withdrawn.

The Commission’s letter has been quoted in the component sections above, and its third objection deserves separate emphasis because it concerns component four rather than disclosure alone. By deleting the ceiling in section 182 and the duty to name recipients, the letter warned, the Act enabled even companies without distributable profits to donate, and opened the door to shell companies created for the sole purpose of channelling funds to parties. The Commission was not a political opponent of the government. It was the body that would have to police section 29B’s prohibitions with the evidence the amendments destroyed, writing to the law ministry of the same government, in the ordinary channel of inter institutional consultation, before any election had been fought under the scheme. When the Commission’s counter affidavit of 27 March 2019 repeated these objections in court, the government responded that the Commission had earlier supported aspects of the proposal in meetings, and the affidavit’s annexed correspondence settled the argument in the Commission’s favour. The documentary spine of the petitioners’ case, in other words, was not journalism or advocacy research. It was the state’s own file notings, produced by the two institutions, the central bank and the election authority, whose statutory duties the scheme most directly engaged.

The formal petitioners assembled around that spine. The Association for Democratic Reforms, the election watchdog whose litigation had produced the disclosure judgments of 2002, filed its writ petition in 2017 challenging the Finance Act amendments, joined by Common Cause, the public interest organisation whose 1996 judgment had disciplined party accounts. The Communist Party of India (Marxist) petitioned as a political party, giving the challenge the unusual feature of a registered party attacking a funding channel available to itself, on the stated ground that the channel’s asymmetry and opacity injured the electoral process as a whole. Jaya Thakur, a Congress office holder, filed a further petition. The petitioners’ grounds tracked the architecture. Article 19(1)(a), because anonymity extinguished the voter’s information about party finance. Article 14, because the scheme’s design favoured the incumbent and treated corporate and individual political money unequally without adequate reason. The Money Bill objection, because four statutes had been amended through a legislative vehicle that denied the Rajya Sabha its revising role. And the statutory ground that the scheme as operated defeated the prohibitions of section 29B and the purposes of the Companies Act amendments it displaced. The Supreme Court heard stay applications in 2019 and again in March 2021, when a bench led by Chief Justice S. A. Bobde declined interim relief ahead of state elections, reasoning that the scheme had operated through earlier elections and that the balance of convenience did not favour a halt. The petitions then waited, while the windows opened and closed and the totals mounted, until the autumn of 2023.

Six Years in Court: Interim Orders, Sealed Covers and a Late Bench

The litigation’s chronology is itself evidence about the architecture, because the scheme’s design features shaped the pace at which the law could examine it. The petitions of 2017 attacked a statute before the notification existed. Once sales began in March 2018, the petitioners sought a stay, and on 12 April 2019, in the middle of the general election campaign, a bench presided over by Chief Justice Ranjan Gogoi passed the interim order that defined the scheme’s judicial existence for five years. The Court declined to stay the scheme, holding that the competing considerations required deeper examination, but it directed every party that had received bond donations to submit to the Election Commission, in sealed cover, the particulars of donors, the amounts received against each bond, and the bank accounts into which the sums had been credited, and it directed the Finance Ministry to reduce the sale window for the election period. The order thus created the paradox of the scheme’s middle years. The information whose suppression the petitioners challenged continued to be generated, continued to be secret from the public, and yet existed in deposited form with the election authority, under the Court’s seal. Anonymity persisted as against the voter while lapsing as against two institutions, the bank and the Commission, which between them held the complete picture the voter was denied.

The sealed cover device attracted criticism from transparency advocates, who noted that evidence deposited secretly in a case about secrecy compounded the original wrong, and from the petitioners themselves when they returned to court. In January 2020 the Association for Democratic Reforms applied for a stay ahead of a notified sale window, arguing that the Commission and the Reserve Bank had both opposed the scheme on record and that each fresh window added sums the voter could not trace. The application did not produce a stay. In March 2021, with elections due in four states and a union territory, the Association sought interim relief again, and the bench of Chief Justice Bobde refused it on 26 March 2021, observing that bonds had been released in earlier years without any suggestion that they had distorted those elections in a manner the Court could act upon, and that the apprehension of foreign money entering through the channel was met by the know your customer requirement at purchase. Sales continued through 2021, 2022 and 2023. The sums involved grew with each year, so that by the time the merits were finally heard, the channel had carried the great majority of the 16,518 crore rupees it would ever carry. Judicial time and sale windows ran on different clocks, and the architecture’s clock was faster.

While the petitions waited, the only public evidence of the scheme’s operation accumulated in the aggregate figures that component two permitted, and those figures were not reassuring to the scheme’s defenders. The contribution reports and audited accounts that parties filed showed bond income concentrating year after year in the ruling party at the centre, with the Bharatiya Janata Party declaring receipts through the channel that exceeded the bond receipts of all other national parties combined in successive financial years, and showed the channel supplying a rising share of the declared income of the parties that used it. The Association for Democratic Reforms published annual compilations of these filings, and the compilations became the petitioners’ empirical brief in waiting: if the channel’s yield divided so unevenly under conditions of secrecy, the argument ran, the secrecy was not neutral between competitors, whatever its formal symmetry. The Union’s answer, that any party was free to cultivate donors and that fundraising success is not a constitutional injury, was the answer of a party system speaking. The petitioners’ reply was that a state designed channel, operated by a state owned bank under a statute the incumbent had enacted through a Money Bill, could not disclaim responsibility for its own gradient as though the gradient were weather.

The merits were reached in the autumn of 2023 for a reason that belongs to the institutional history this pillar records. On 16 October 2023, with a general election months away and the petitioners pressing that a decision after the election would be a decision after the money had been spent, Chief Justice Chandrachud directed that the petitions be listed before a five judge Constitution Bench on 31 October 2023. The hearings ran for three days, 31 October, 1 November and 2 November 2023, and the bench reserved judgment on 2 November. The composition of the bench is part of the record: the Chief Justice, Justice Sanjiv Khanna, Justice B. R. Gavai, Justice J. B. Pardiwala and Justice Manoj Misra. The arguments reprised the architecture in adversarial form. For the petitioners, counsel including Prashant Bhushan and Kapil Sibal pressed the Article 19(1)(a) line from the 2002 and 2003 judgments, the asymmetry argument from the disclosed aggregate figures, which already showed the ruling party receiving a majority of all bond income, and the section 182 objection that unlimited and unnamed corporate money had converted company law into an electoral weapon. For the Union, the Attorney General and the Solicitor General defended donor privacy, the banking channel’s discipline, and legislative primacy in economic policy. The Election Commission, consistent with its letter of 2017 and its affidavit of 2019, told the Court that it stood for transparency and disclosure, a position the final judgment would note with care. Readers who want the institutional frame for the Commission’s conduct across these years will find it in the institution-under-stress framework, which traces how the Commission’s stated positions have fared when they collide with the executive that appoints its members.

The Judgment of 15 February 2024, Stated by Holding

The judgment in Association for Democratic Reforms v Union of India was delivered on 15 February 2024, and it was unanimous in result. Two opinions compose it: the lead opinion of Chief Justice Chandrachud, written for himself and for Justices Gavai, Pardiwala and Misra, and a concurring opinion of Justice Khanna, who reached the same conclusions by a proportionality analysis he described as carrying a slight variation. The holdings are stated here as the Court stated them, because the architecture reading depends on their exact scope.

The lead opinion held, first, that information about the funding of a political party is essential for a voter to exercise the freedom to vote in an effective manner, that the right to information under Article 19(1)(a) therefore extends to the sources of party finance, and that the Electoral Bond Scheme and the amended provisions, to the extent that they anonymised contributions through bonds, violated Article 19(1)(a). The reasoning built the holding in stages. Economic inequality in a polity translates into political inequality, because the contributor of large sums obtains a seat at the table where policy is shaped and obtains access that the ordinary voter cannot match. Contributions of scale are made, in the ordinary course, with the expectation of a return, and the judgment used the term quid pro quo for arrangements in which policy, contracts or licences answer donations. A voter assessing a party’s programme is entitled to know whether interests financing the party may be shaping the programme, since that knowledge bears directly on the choice the ballot records. The Court rejected the Union’s submission that the right to information stops at the candidate, holding that the voter’s informational entitlement recognised in 2002 and 2003 was grounded in the effective exercise of the franchise and could not be confined to one category of electorally material fact.

The lead opinion held, second, that the anonymity the scheme conferred was not justified as a proportionate restriction. Applying the proportionality enquiry, the Court accepted that curbing unaccounted cash in political funding was a legitimate aim and that the donor’s interest in privacy of political affiliation was a real interest, but held that the scheme’s means bore no adequate proportion to those aims. The instrument applied anonymity without limit of amount, so that a gift of one crore rupees and a gift of one thousand rupees received identical concealment, although the privacy interest and the influence interest differ across that range by orders of magnitude. Less restrictive alternatives existed and had been placed before the Court, including the electoral trust model with its named reporting and the option of disclosure above a threshold. A measure that protects the small donor’s privacy by extinguishing the voter’s knowledge of the largest contributions in the system failed the balancing stage of the enquiry, at which the Court also weighed the scheme’s demonstrated incapacity to perform its stated purpose: anonymous banked money had not displaced cash, which continued to circulate in elections in quantities the Commission’s own seizures documented year after year.

The lead opinion held, third, that the amendment to section 182 of the Companies Act was manifestly arbitrary and violated Article 14. The deletion of the profit linked ceiling permitted contributions unconnected to a company’s commercial substance, and the bench reasoned that a company that has made no profits, or losses, can contribute only from capital or borrowings, which is to say from other people’s money, in pursuit of returns that are not commercial in any ordinary sense. The amendment also obliterated a constitutionally relevant distinction by placing corporate contributors and individual contributors under a single permissive rule, ignoring that a company’s contribution is decided by those who control it while its costs fall on shareholders who may hold opposite political convictions, and that the corporate form can be manufactured at will for the purpose of giving. Unlimited corporate contributions, the Court concluded, injure the principle of free and fair elections, itself a basic feature, by amplifying the political voice of concentrated capital beyond any proportion to the voices of citizens.

Fourth, the Court disposed of the remaining defences. It declined to treat the scheme as unreviewable economic policy, holding that a measure structuring the flow of money into elections touches the electoral process directly and is amenable to fundamental rights review. It held the amendments to section 29C and to section 13A unconstitutional to the extent of the anonymity they conferred, rather than as free standing wrongs, which preserved the reporting statutes in their unamended reach. It did not find it necessary to rest the result on the Money Bill objection, the issue of the Finance Act’s certification having been referred in a separate line of cases. The operative directions were immediate. The State Bank was ordered to stop issuing bonds forthwith. The bank was ordered to furnish to the Election Commission, by 6 March 2024, the details of every bond purchased since the interim order of 12 April 2019, including the date of purchase, the name of the purchaser and the denomination, together with the details of every bond encashed by each party. The Commission was ordered to publish the information on its official website by 13 March 2024. Bonds within their fifteen day validity that had not been encashed were to be returned and refunded to the purchasers’ accounts.

What directions did the Court give the State Bank on 15 February 2024?

The Court directed the bank to stop issuing bonds at once, to deliver to the Election Commission by 6 March 2024 the date, purchaser name and denomination of every bond bought since 12 April 2019 and the encashment record of every party, and to refund unencashed bonds still inside their validity.

The Concurrence: Justice Khanna and the Shape of Unanimity

Justice Khanna’s separate opinion is frequently compressed into a footnote in commentary, and the compression loses something the architecture reading needs. Justice Khanna agreed with every conclusion and every direction in the lead opinion, and said so. His variation lay in method. Where the lead opinion applied what it called a double proportionality enquiry, weighing the voter’s right against the donor’s privacy claim in a structured balance, Justice Khanna applied the proportionality test in its standard form and supplemented it with a close examination of the data the State Bank had furnished to the Court during the hearings, including the distribution of purchases across denominations and the pattern of contributions. On that examination he recorded that the bulk of bonds by value had been purchased at the highest denomination, a datum he treated as material to the balance, because an instrument justified as protection for the modest donor was in practice an instrument of the largest givers, and a justification that does not match the use to which a measure is put cannot carry the measure through the necessity stage of proportionality. He also addressed the privacy claim at its root, accepting that political affiliation and preference lie close to the core of the privacy that the Constitution protects after the Puttaswamy judgment of 2017, yet holding that a donation to a political party is not a private act of conscience alone. It is an act with public consequences, performed in the arena the franchise governs, and the privacy that protects a citizen’s vote, cast in secret by the design of the ballot itself, does not extend to a citizen’s purchase of influence at wholesale denominations.

Which earlier judgments supplied the right to information about party finance?

The lead opinion built on Union of India v Association for Democratic Reforms of 2 May 2002 and the People’s Union for Civil Liberties judgment of 13 March 2003, which grounded the voter’s Article 19(1)(a) right to electorally material information in the effective exercise of the franchise.

The fact of unanimity deserves one paragraph of institutional attention. Constitution Benches in politically charged cases have fractured along lines commentators predicted from the appointing governments of their members, and the electoral bonds bench did not. A Chief Justice appointed in the ordinary course of seniority, sitting with colleagues whose elevations spanned different administrations, produced a five judge consensus that a signature policy of the incumbent government violated the Constitution, and produced it in an election season, with directions timed to place the withheld information before the electorate ahead of the poll. Whatever account a reader takes of the judgment’s doctrine, the event belongs in the record of the institution-under-stress framework as a counterweight case, in which a constitutional institution under pressure discharged the function the pressure was meant to deter. The comparison with the election technology cases is instructive here, and the technological-architectural-trajectory comparator shows a bench culture wrestling with a different machine of electoral administration in the same period, deferring where the bonds bench intervened, which suggests the variable was not judicial appetite in the abstract but the quality of the right asserted and the completeness of the documentary record the objecting institutions had built.

The Disclosure of March 2024: The Architecture Run in Reverse

The judgment’s directions set a timetable, and the timetable nearly failed at its first step. On 4 March 2024 the State Bank applied to the Supreme Court for an extension of time to 30 June 2024, pleading that decoding the siloed data of purchases and encashments and matching the two across its branches required months of work. The Association for Democratic Reforms answered with a contempt application pointing out that the bank had maintained the data in digital form throughout and that the requested date fell after the general election, which would postpone the voter’s knowledge until the vote it was meant to inform had been cast. On 11 March 2024 the Constitution Bench dismissed the extension application, recorded its displeasure at the plea, and directed the bank to comply by the close of business on 12 March, warning that failure would expose it to contempt. The bank transmitted the data on 12 March. On 14 March 2024 the Election Commission published the datasets on its website, on the as is where is basis on which it had received them, reminding readers in its covering statement that it had consistently and categorically weighed in favour of disclosure and transparency, a claim its letter of 26 May 2017 substantiated. One defect remained. The published files listed purchasers with dates and denominations and parties with dates and denominations, but without the unique alphanumeric number printed on each bond, the two lists could not be joined. On 18 March 2024 the Court, hearing the matter again, directed the bank to disclose the bond numbers and everything else in its possession, observing that the judgment had required complete disclosure. The bank’s chairman filed an affidavit of compliance, and on 21 March 2024 the complete data, bond numbers included, was published. Seven years of statutory anonymity ended in a fortnight of compelled publication.

What the record showed, once journalists and researchers had joined the lists by number and date, can be organised under three heads. The first is scale. Across the scheme’s whole life, from March 2018 to the final phase of January 2024, the State Bank sold 28,030 bonds worth 16,518.11 crore rupees, a sum exceeding the declared election expenditure of every party in the system combined across several cycles. The disclosed window, purchases from 12 April 2019 onward, accounted for the great majority of that value. The second head is concentration on the receiving side. Party declarations across the scheme’s life, compiled by the Association for Democratic Reforms from the contribution and audit filings, credited the Bharatiya Janata Party with 6,566 crore rupees, some fifty five percent of all bond money encashed, against 1,123 crore rupees for the Indian National Congress and 1,092 crore rupees for the All India Trinamool Congress. In the court ordered disclosed set, matched bond by bond after 21 March, the same ordering appeared with sharper edges: the Bharatiya Janata Party encashed 6,060.5 crore rupees, about forty seven and a half percent of the redeemed value in that set, followed by the Trinamool Congress at 1,609.5 crore rupees, the Congress at 1,421.9 crore rupees, the Bharat Rashtra Samithi at roughly 1,214 crore rupees, the Biju Janata Dal at 775 crore rupees and the Dravida Munnetra Kazhagam at about 639 crore rupees. Parties governing states drew sums out of proportion to their national footprints, which the data made visible as a structural feature rather than a partisan observation: control of a government, at either level, was the strongest predictor of bond receipts, because the buyers were purchasing proximity to decisions that governments take.

The third head is the purchasing side, where the architecture’s predictions met their exhibits. The largest single purchaser was Future Gaming and Hotel Services Private Limited, the lottery company associated with Santiago Martin, which bought 1,208 bonds worth 1,368 crore rupees across the disclosed period, a company whose bank accounts the Enforcement Directorate had attached in March 2022 in a money laundering investigation and against which a prosecution complaint stood filed by September 2023. The second purchaser was Megha Engineering and Infrastructures Limited of Hyderabad, at 966 crore rupees, a contractor whose order book rested substantially on government infrastructure awards. Behind them came Qwik Supply Chain Private Limited at 410 crore rupees, a company whose purchases of 360 crore rupees in the financial year 2021-22 stood against a recorded net profit of 21.72 crore rupees for that year, Vedanta Limited at 400 crore rupees, Haldia Energy Limited at 377 crore rupees, and a roll of twenty two firms in all that purchased above one hundred crore rupees each. Journalists at the Reporters’ Collective, Newslaundry and the Hindu, working the matched datasets in the weeks after 21 March, documented the temporal patterns the petitioners had predicted in the abstract: purchases clustered in the days after raids, searches or summonses by central investigative agencies against the purchasing companies, purchases preceding the award of contracts to the purchaser, and donations by companies that were loss making in the years of their giving. The companies and parties concerned denied any quid pro quo when asked, and no court has adjudicated those individual transactions. The patterns are recorded here as documented patterns in a published record, contested by their subjects, because that is their evidentiary status. Their analytical status is different, and stronger. Component four had deleted the profit tether, component one had hidden the purchaser, and the disclosed data showed transactions that only the combination of those two deletions made possible: giving untethered from commercial substance, at wholesale denominations, in temporal proximity to the exercise of state power over the giver.

The disclosure also carried a boundary that its critics marked at once, and the boundary belongs in the record because it qualifies every total quoted from the published data. The judgment’s directions ran from the interim order of 12 April 2019, the date from which the Court had compelled the deposit of particulars, and the published datasets accordingly began at that date. The scheme’s first year, from March 2018 to April 2019, stood outside the court ordered window. The State Bank’s own figures, cited in a subsequent application by an intervenor seeking disclosure of the complete record, put the undisclosed residue at 9,159 bonds worth 4,002 crore rupees, some twenty four percent of the value the channel ever carried, purchased during a period that included the run up to the 2019 general election. The applicant argued that a scheme held violative of Article 19(1)(a) could not leave a quarter of its transactions in darkness by reason of a date chosen for an interim purpose. The argument states the boundary’s oddity fairly. What the country received in March 2024 was the architecture’s record for five of its six years, complete within those years and silent about the first, and every analysis built on the matched data, including the analysis in this article, inherits that silence.

One further finding of the disclosure season bears on the framing dispute this article must adjudicate. The aggregate receipts that parties had published during the scheme’s life, the totals component two permitted, had created an impression in commentary that the country knew the broad shape of bond finance and lacked only the names. The matched data showed the impression was false in a way totals could never reveal. The shape itself had been invisible: the dominance of the single crore denomination, the clustering of purchases around agency actions and contract decisions, the presence of purchasers with no profits to give from, the identity of the largest donor as a company under active money laundering prosecution. Composition was not a refinement of the total. It was the substance. The point had been made in doctrine by the lead judgment in February, when it held that information about funding is essential to the voter’s effective choice. The data of March performed the same point empirically, and the interval between the two months is the interval between a right asserted and a right exercised. Readers tracing the election that this finance underwrote can set the disclosure beside the post-2019 election political-architectural context, where the aggregate figures available at the time are the figures component two allowed the country to see.

Three Readings of the Trajectory, and a Judgment Between Them

Scholarship and commentary on the bonds have organised themselves into two readings, and the brief for this article requires a judgment between them and a third. The first reading treats the scheme as political finance reform that failed on its merits. Its most careful exponent in the legislative record is M. R. Madhavan of PRS Legislative Research, whose analyses of the Finance Bill 2017 and of the scheme’s operation, written as the events unfolded, catalogued the departures from the disclosure regime, questioned the use of a Money Bill to amend four statutes, and measured the scheme against the reform consensus that had accumulated since the Indrajit Gupta Committee. On this reading the bonds are a chapter in a longer reform story that begins with the Rao government’s institutional-political context, where liberalisation multiplied the points at which money and policy meet, and the chapter’s moral is that a reform which moves against the direction of two decades of expert consensus should be expected to end as this one ended. The second reading treats February 2024 as the story. Suhrith Parthasarathy’s essays in the Economic and Political Weekly on the verdict exemplify it: the judgment as a landmark in the jurisprudence of Article 19(1)(a), the restoration of the voter’s informational right after a seven year eclipse, the unanimous bench as institutional redemption. Niranjan Sahoo’s essays on Indian political finance, written from the Observer Research Foundation across the scheme’s life, supply the second reading’s policy tail, arguing that the verdict returns the country to the unresolved agenda of money power, where state funding, expenditure regulation and party law reform wait where the Indrajit Gupta and Law Commission reports left them.

The two readings also divide the scholarly labour in a way worth noticing. Madhavan’s PRS analyses belong to the genre of legislative scrutiny, written while the bill and then the scheme could still be amended, and their characteristic question is the drafter’s question: what does this clause do to the disclosure regime the country already has? Parthasarathy’s essays belong to the genre of constitutional commentary, written after judgment, and their characteristic question is the doctrine’s question: what does this holding do to the Article 19(1)(a) tradition the Court has built since 2002? Sahoo’s work moves between the genres, measuring each legal event against the structural variable he treats as decisive, the sheer quantum of money Indian elections absorb and the absence of any lawful supply adequate to it. A reader equipped with all three genres can describe the episode completely at the level of law, doctrine and political economy, and still miss the level at which the episode was actually transacted, the level of institutional design, where a bank’s ledger rules, a reporting proviso, a denomination table and a company law ceiling were assembled into a single functioning machine. That is the level the architecture reading claims, and the claim is empirical before it is rhetorical: the machine’s parts are documents, its operations are dated, and its output, 16,518 crore rupees of unattributed party finance, is a matter of published record.

Each reading captures something the record supports, and each collapses something the record contains. The reform reading cannot explain the design. A reformer who misjudges one feature does not misjudge all four in the same direction, deleting the disclosure rule, the record rule, the ceiling and the naming rule while choosing the bearer form against the central bank’s written advice, each choice subtracting information or adding volume. The pattern of choices is the architecture, and a framing that evaluates the scheme as an attempt at transparency that stumbled cannot say why every stumble landed on the same side. The verdict reading, for its part, cannot explain the disclosure. If the story were the judgment, March 2024 would be an epilogue, the mechanical execution of a direction. In fact the disclosure was the moment the scheme’s concealed substance entered public knowledge, the moment at which claims made in affidavits since 2017 became checkable facts, and the moment the country’s understanding of its own elections altered. A framing that centres the courtroom treats the most consequential month of the episode as administration.

The anonymity architecture reading, defended here, holds that the analytical content of the episode is the design and its unmaking. It explains the reform reading’s anomaly, because a system whose joint product is anonymity predicts exactly the pattern of choices the record shows. It explains the verdict’s actual shape, because a bench dismantling an architecture must strike component by component, the scheme, the reporting exemptions and the Companies Act amendment together, which is what the twin opinions did. It explains the disclosure’s force, because disclosure was the architecture run in reverse, the bank’s ledger published and the contribution report the statute had forbidden reconstructed by court order. And it connects the episode to the series’ larger frame. Read through the integrated three-tensions framework, the bonds appear as a renegotiation of the founding settlement on political finance, the 1951 assumption that party money would be modest, member raised and visible, replaced by a settlement in which party money was corporate, wholesale and invisible, until a court restored the older settlement’s informational premise on constitutional grounds. The reading also survives the strongest counter against it, which must be stated plainly. The counter holds that the architecture framing, absorbed in one machine, under engages the durable question the machine was built to answer: how a democracy finances expensive elections without either cash corruption or the exposure of donors to retaliation. That question survived the verdict intact, and the final section takes it up on its merits rather than treating the strike down as the end of thought.

The Question the Verdict Left Standing

The judgment of 15 February 2024 destroyed an architecture and restored a legal position. With the amendments struck down, section 182 returned to its 2013 form, ceiling and naming duty included, and sections 29C and 13A again governed bond like contributions as they govern every contribution, there being no bonds left to exempt. The general election of April to June 2024 was accordingly the first since 2014 to be fought without the instrument, financed through the older channels of cheque, transfer, electoral trust and cash below the 2,000 rupee anonymity limit. Nothing replaced the scheme. No successor statute was introduced in the period the record covers, and the finance of the 2024 campaign ran through instruments whose disclosure rules predate the episode, including the electoral trusts whose donor reporting the 2024 data showed to be the road not taken in 2017. The restoration is therefore double edged for every camp in the argument. The petitioners recovered the law they had defended, which is the law under which sixty nine percent of party income had arrived from unknown sources in the decade before the bonds. The government’s diagnosis of that older regime was never refuted by the judgment, because the judgment did not need to refute it. The Court held that the cure was unconstitutional, not that the disease was imaginary.

What remains open, and what the architecture reading obliges its adherents to face, is the design problem in its honest form. Donor retaliation is a documented feature of polities where the state is the largest buyer, licensor and lender, and the Indian state’s own agencies supplied the disclosed data’s most uncomfortable exhibits, of purchases following enforcement actions. A disclosure regime without any protection invites the cash economy back, as the budget speech warned. A protection regime without disclosure invites the bond economy back, as the six years demonstrated. The instruments that other democracies and India’s own reform literature have placed between those poles are known and dated: disclosure above a threshold set low enough to catch influence and high enough to shelter the small giver, state funding in kind of the sort the Indrajit Gupta Committee recommended in 1998, expenditure ceilings enforced at the level of the party rather than the candidate alone, and an election authority with the staff and the statutory warrant to audit party accounts in real time rather than in retrospect. Each has a literature, named authors and recorded objections, and none was enacted in the wake of the verdict within the period this article surveys. The episode’s final lesson is therefore institutional rather than doctrinal. Between 2017 and 2024 the Indian state built, defended and lost an anonymity architecture, and the record of the building, preserved in the Reserve Bank’s letters, the Election Commission’s objections, the sealed covers and the published datasets, constitutes the most complete evidentiary base any future designer of Indian political finance will possess. The journalists who made that record usable, joining bank lists to party lists while the sources who explained the files guarded their own names against the retaliation the case was about, practised the discipline described in accounts of how journalists use VaultBook to protect sources, and their work is part of the record too. The architecture is gone. The question it answered badly remains, awaiting an answer built to be looked at.

Frequently Asked Questions

Q: What were electoral bonds?

Electoral bonds were bearer instruments in the nature of promissory notes, interest free and payable to the bearer on demand, which the State Bank of India alone was authorised to issue under the Electoral Bond Scheme notified on 2 January 2018. A buyer purchased a bond through a know your customer compliant bank account, by cheque or electronic transfer, and delivered the physical instrument to a political party, which deposited it in a designated account and received the face value. The bonds came in denominations of 1,000 rupees, 10,000 rupees, one lakh rupees, ten lakh rupees and one crore rupees, carried no name, and expired fifteen calendar days after issue. The scheme operated through thirty sale phases between March 2018 and January 2024, moving 16,518.11 crore rupees, before the Supreme Court struck it down on 15 February 2024.

Q: When was the electoral bond scheme introduced?

The scheme was announced by Finance Minister Arun Jaitley in the Union Budget speech of 1 February 2017, enacted through the Finance Act 2017, which amended the Reserve Bank of India Act, the Representation of the People Act, the Income Tax Act and the Companies Act, and operationalised by the Electoral Bond Scheme notification published in the Gazette on 2 January 2018. The first bonds were sold in March 2018 through four authorised branches of the State Bank of India, a network that grew to twenty nine branches. The Finance Act 2017 also cut the ceiling for anonymous cash donations to parties from 20,000 rupees to 2,000 rupees, so the bond opened as an unlimited anonymous banking channel in the same statute that tightened the anonymous cash channel, a pairing that shaped the constitutional litigation from 2017 to the judgment of 15 February 2024.

Q: Who could buy electoral bonds?

Any citizen of India, and any entity incorporated or established in India, could purchase electoral bonds, provided the purchase was made from a know your customer compliant account with the State Bank of India and paid for by cheque or electronic transfer rather than cash. Persons could buy singly or jointly with others. The receiving side was narrower than the buying side. Only a political party registered under section 29A of the Representation of the People Act 1951 that had secured at least one percent of the votes polled at the most recent general election to the Lok Sabha or to a state legislative assembly could encash a bond, and encashment had to occur through the party’s designated account within the instrument’s fifteen day validity. The Reserve Bank of India had warned in January 2017 that the bearer form meant the final contributor of a transferred bond could not be known even where the original purchaser was recorded.

Q: What did the February 2024 verdict say?

In Association for Democratic Reforms v Union of India, delivered on 15 February 2024, a five judge Constitution Bench presided over by Chief Justice D. Y. Chandrachud, with Justices Sanjiv Khanna, B. R. Gavai, J. B. Pardiwala and Manoj Misra, unanimously struck down the Electoral Bond Scheme. The lead opinion held that information about the funding of a political party is essential for a voter to exercise the freedom to vote effectively, that the scheme and the amended provisions of the Representation of the People Act and the Income Tax Act violated the right to information under Article 19(1)(a) to the extent that they anonymised contributions, and that the amendment to section 182 of the Companies Act was manifestly arbitrary. The Court directed the State Bank of India to stop issuing bonds and to disclose the purchase and encashment record since 12 April 2019 to the Election Commission for publication.

Q: What did the SBI data reveal?

The data the State Bank of India transmitted on 12 March 2024 and the Election Commission published on 14 March 2024, completed by the bond numbers disclosed on 21 March 2024, revealed the composition that six years of aggregate totals had concealed. Purchases across the scheme’s life totalled 28,030 bonds worth 16,518.11 crore rupees. The largest purchaser was Future Gaming and Hotel Services, at 1,368 crore rupees, followed by Megha Engineering at 966 crore rupees and Qwik Supply Chain at 410 crore rupees. The Bharatiya Janata Party was the largest recipient, encashing 6,060.5 crore rupees in the disclosed set, ahead of the Trinamool Congress and the Congress. The matched record also showed purchases clustered at the one crore denomination, purchases by companies whose giving exceeded their recorded profits, and purchase dates that journalists set against investigative agency actions and government contract awards affecting the purchasers.

Q: Why were electoral bonds anonymous?

Anonymity was the scheme’s stated inducement, not its accident. In the budget speech of 1 February 2017, Arun Jaitley argued that donors feared adverse consequences if their political giving were disclosed, since governments control contracts, licences and inspections, and that this fear drove political money into untraceable cash. The bond offered a bargain: money would move through taxed bank accounts, whose trail the state could reconstruct, while the donor’s name would be recorded only by the issuing bank and disclosed to no voter, party rival or election authority. The Finance Act 2017 hard wired the bargain by exempting bond contributions from the contribution report under section 29C of the Representation of the People Act and from the donor record duty under section 13A of the Income Tax Act. The Supreme Court held in February 2024 that the bargain extinguished the voter’s Article 19(1)(a) right to information about party finance and that less restrictive means of protecting donors existed.

Q: What is Section 182 of the Companies Act?

Section 182 of the Companies Act 2013 is the provision governing political contributions by companies. As enacted, it permitted a company to contribute up to seven and a half percent of its average net profits during the three preceding financial years, required authorisation by a resolution of the board of directors, and required disclosure in the profit and loss account of the total amount contributed and the names of the recipient parties. The Finance Act 2017 deleted the profit linked ceiling and deleted the duty to name recipient parties, leaving only an aggregate figure, so that a company could contribute sums exceeding its profits and report the outgo without attribution. The Election Commission warned in May 2017 that the change invited shell companies formed solely to channel donations. On 15 February 2024 the Supreme Court held the amendment to section 182 manifestly arbitrary and unconstitutional, restoring the ceiling and the naming duty of the 2013 text.

Q: On what grounds did the petitioners challenge the scheme?

The Association for Democratic Reforms, Common Cause, the Communist Party of India (Marxist) and Congress office holder Jaya Thakur challenged the scheme on connected grounds. Under Article 19(1)(a), they argued that anonymity extinguished the voter’s right to information about the funding of political parties, a right the Supreme Court had recognised in the candidate disclosure judgments of 2002 and 2003. Under Article 14, they argued that the design favoured the incumbent party, which could raise the most through the channel and command the institutions able to pierce its secrecy, and that unlimited corporate contributions stood on no rational footing beside individual giving. They attacked the use of a Money Bill to amend four statutes, and they argued that the exemptions defeated section 29B of the Representation of the People Act, which prohibits parties from accepting government company and foreign donations that could no longer be detected.

Q: What did the Reserve Bank of India warn in January 2017?

When the Finance Ministry sought its comments on 28 January 2017, the Reserve Bank of India replied on 30 January 2017 under Governor Urjit Patel opposing the design. The bank warned that electoral bonds would be bearer instruments transferable by delivery, so that who finally contributed such an instrument to a party would not be known; that bearer instruments have the potential to become currency; that issuing them in sizeable quantities could undermine faith in Indian banknotes; and that the amendment to the Reserve Bank of India Act would set a bad precedent by encouraging money laundering and eroding a core principle of central banking legislation. The bank’s committee of the central board urged that any such instrument be issued in electronic, dematerialised form to reduce misuse. The Finance Ministry retained the paper bearer form, and the correspondence, obtained under the Right to Information Act in 2019, became part of the petitioners’ record in the Supreme Court.

Q: What did the interim order of 12 April 2019 require?

The interim order of 12 April 2019, passed by a bench presided over by Chief Justice Ranjan Gogoi in the middle of the general election campaign, declined to stay the Electoral Bond Scheme, holding that the competing contentions required deeper examination. It nevertheless built the evidentiary bridge to the disclosure of 2024. Every political party that had received donations through bonds was directed to submit to the Election Commission, in sealed cover, the particulars of its donors, the amounts received against each bond and the credit account details into which the sums were deposited, and the Finance Ministry was directed to reduce the sale window then open. The order meant that the record of the scheme’s operation continued to be generated and deposited with the election authority even while it remained hidden from the public, and the judgment of 15 February 2024 later fixed 12 April 2019 as the date from which the State Bank’s purchase and encashment record had to be disclosed and published.

Q: How much money passed through electoral bonds in total?

Across the scheme’s operating life, from the first sale phase of March 2018 to the final phase of January 2024, the State Bank of India sold 28,030 electoral bonds worth 16,518.11 crore rupees, according to the bank’s figures compiled by the Association for Democratic Reforms. The sum exceeded the declared income most parties reported from all sources across several of those years, and it arrived through thirty sale phases of ten days each, with longer windows in the general election year. Party filings across the whole period credited the Bharatiya Janata Party with 6,566 crore rupees, roughly fifty five percent of all bond money encashed, the Congress with 1,123 crore rupees and the Trinamool Congress with 1,092 crore rupees. The court ordered disclosure of March 2024 covered purchases from 12 April 2019 onward, the great majority of the total value, and allowed those aggregates to be decomposed donor by donor and party by party for the first time.

Q: Which companies bought the largest value of bonds?

The disclosed State Bank record identified Future Gaming and Hotel Services Private Limited, the lottery company associated with Santiago Martin, as the largest purchaser, with 1,208 bonds worth 1,368 crore rupees bought between April 2019 and January 2024, a period during which the Enforcement Directorate attached its bank accounts in a money laundering investigation. Megha Engineering and Infrastructures Limited of Hyderabad followed at 966 crore rupees, then Qwik Supply Chain Private Limited at 410 crore rupees, Vedanta Limited at 400 crore rupees, Haldia Energy Limited at 377 crore rupees, the Bharti group at 247 crore rupees, Essel Mining and Industries at 224 crore rupees and Western UP Power Transmission at 220 crore rupees. Twenty two firms in all purchased more than one hundred crore rupees each. Individual buyers also appeared in the record, including industrialists purchasing in their own names, but corporate purchases dominated the value of the channel from its first phase to its last.

Q: What share of the bond money did the largest recipient obtain?

By every measure the record supplies, the Bharatiya Janata Party obtained the largest share. Party declarations across the scheme’s whole life, compiled by the Association for Democratic Reforms from filings with the Election Commission, credited the party with 6,566 crore rupees, about fifty five percent of all bond money encashed between March 2018 and January 2024, against roughly nine percent each for the Congress and the Trinamool Congress. In the narrower disclosed set that the Supreme Court ordered published, bonds purchased from 12 April 2019 onward and matched after the bond numbers were released on 21 March 2024, the party encashed 6,060.5 crore rupees, about forty seven and a half percent of the redeemed value in that set, followed by the Trinamool Congress at 1,609.5 crore rupees and the Congress at 1,421.9 crore rupees. State governing parties of several formations also drew sums large relative to their size, a pattern the judgment’s analysis of incumbency advantage had anticipated.

Q: Why did Justice Khanna write a separate concurring opinion?

Justice Sanjiv Khanna agreed with every conclusion and direction in the lead opinion of Chief Justice Chandrachud, and his concurrence is a variation in method rather than a disagreement in result. Where the lead opinion applied a double proportionality enquiry to the competing claims of the voter’s right to information and the donor’s privacy, Justice Khanna applied the proportionality test in its standard form and supplemented it with a close reading of the purchase data the State Bank of India had placed before the Court, recording that the bulk of value had been bought at the one crore denomination. That distribution mattered to his balancing, because an instrument defended as protection for the modest donor operated in practice as an instrument of the largest givers. He also held that a donation is a public act with public consequences, distinguishable from the secrecy of the ballot itself, so that the privacy protecting a citizen’s vote could not be extended to wholesale purchases of political access.

Q: What happened when the State Bank sought more time in March 2024?

On 4 March 2024 the State Bank of India applied to the Supreme Court for an extension of the disclosure deadline to 30 June 2024, pleading that matching its siloed purchase and encashment data across branches required months of work. The Association for Democratic Reforms answered with a contempt application, noting that the requested date fell after the general election and would postpone the voter’s knowledge until after the vote the information was meant to inform. On 11 March 2024 the Constitution Bench dismissed the extension application, recorded that the data had been maintained in digital form, and directed compliance by the close of business on 12 March 2024, warning that failure would invite contempt proceedings. The bank transmitted the data on 12 March, and the Election Commission published it on its website on 14 March 2024, two days inside the judgment’s outer date, on the as is where is basis on which it had been received.

Q: How did the bond serial numbers complete the disclosure in March 2024?

The datasets published on 14 March 2024 came in two unjoined halves: a list of purchasers with dates and denominations, and a list of parties with encashment dates and denominations. Without the unique alphanumeric number printed on each bond, no reader could establish which purchaser’s money had reached which party, and the disclosure reproduced, in milder form, the separation the scheme itself had enforced. On 18 March 2024 the Supreme Court directed the State Bank of India to disclose the bond numbers and all other particulars in its possession, observing that the judgment of 15 February had required complete disclosure and that the bank could not select the convenient parts of the record. The bank’s chairman filed an affidavit stating that the complete data had been furnished, and on 21 March 2024 the Election Commission published the numbers. Journalists and researchers could then match purchase to encashment bond by bond, converting two inert lists into the donor to party map that the matched analysis of the following weeks reported.

Q: How does an electoral trust differ from an electoral bond?

An electoral trust is a pooled vehicle, introduced by the Electoral Trusts Scheme of 31 January 2013, under which a registered trust collects contributions from companies and individuals and distributes the pooled funds to political parties. The decisive difference is disclosure. A trust must report to the Election Commission the names of its contributors and the parties to which it distributed funds, and the parties report trust receipts in the ordinary way, so the money is pooled without being anonymised: a voter can learn who gave to the trust and which parties the trust supplied, even though an individual contribution cannot be traced to an individual party rupee for rupee. An electoral bond pooled nothing and disclosed nothing, identifying neither the purchaser to the public nor the composition of a party’s receipts to the Commission. The Supreme Court’s lead opinion of 15 February 2024 noticed the trust model among the less restrictive alternatives that could protect ordinary donors while preserving the voter’s information, which is one reason the scheme failed the proportionality enquiry.

Q: What did the Court say about donations by loss making companies?

The lead opinion treated donations untethered from profit as a symptom of the section 182 amendment’s arbitrariness. Before the Finance Act 2017, a company could contribute at most seven and a half percent of its average net profits across the preceding three financial years, a ceiling that tied political giving to demonstrated commercial substance. The amendment deleted the ceiling, so a company that had earned no profit, or had lost money, could contribute any sum its board approved, necessarily from capital or borrowings rather than earnings. The Court reasoned that contributions of that character are not ordinary corporate expenditure, that they raise the prospect of shell companies created to channel other people’s money into politics, and that the disclosed record bore the apprehension out, with purchases exceeding the recorded profits of the purchasing companies in documented instances. The prospect of a loss making company purchasing political goodwill at scale was, in the bench’s analysis, integral to why the amendment failed Article 14 as manifestly arbitrary.

Q: Did the verdict restore the corporate donation ceiling of the Companies Act 2013?

Yes, in effect. The judgment of 15 February 2024 declared the amendment that the Finance Act 2017 had made to section 182 of the Companies Act 2013 unconstitutional and manifestly arbitrary, and the consequence of striking down an amendment is that the provision stands in its prior form. The ceiling of seven and a half percent of average net profits across the three preceding financial years therefore governs corporate political contributions again, as does the duty to disclose in the profit and loss account the total amount contributed and the names of the parties that received it, the twin requirements the 2017 amendment had deleted. The amendments to section 29C of the Representation of the People Act and section 13A of the Income Tax Act were struck down to the extent that they anonymised contributions, so the reporting of contributions above the statutory threshold and the keeping of donor records again apply to every channel of giving without exception for any instrument.

Q: What replaced the electoral bond scheme after February 2024?

No successor instrument replaced the scheme. The judgment of 15 February 2024 stopped issuance immediately, and no new statute creating an anonymous or semi anonymous funding channel was enacted in the period the record covers. Party finance returned to the channels that predated 2018: donations by cheque and electronic transfer, reported under section 29C of the Representation of the People Act above the statutory threshold; electoral trusts, which report their contributors and their distributions to the Election Commission; and cash donations below the 2,000 rupee anonymity ceiling that the Finance Act 2017 had set and the judgment left untouched. The general election of April to June 2024 was fought under that restored regime, the first general election since 2014 without bonds in circulation. Reform proposals with longer pedigrees, including state funding in kind of the kind the Indrajit Gupta Committee recommended in 1998 and party level expenditure regulation from the Law Commission’s reports, remained proposals, debated in commentary on the verdict rather than enacted in its wake.