When a country of 1.4 billion people sustains GDP growth averaging 6% per year for two decades and still ranks 134th on the Human Development Index (HDI), the gap between economic expansion and human progress demands more than a statistical explanation. It demands a political one [2]. India’s HDI score of 0.644 in 2023 places it alongside nations with per capita incomes less than half its own. Public education spending, at roughly 2.9% of GDP, falls below the target that India’s own National Education Policy 2020 prescribed. Health expenditure, at 2.1%vof GDP, is the second-lowest in the G20 [3]. Learning poverty among rural primary school children, as documented in the Annual Status of Education Reports across an uninterrupted series from 2012 to 2022, has barely shifted in a decade despite nominal increases in school enrolment and aggregate social sector budgets [4].

The conventional explanation for this divergence points to state capacity, colonial institutional legacies, social fragmentation, and the sheer administrative complexity of governing a continental nation where fourteen major languages are spoken across 28 states. These factors are real and consequential. But they do not explain why the composition of public expenditure has moved systematically against long-run investment over precisely the period in which democratic competition has intensified. Nor do they explain why states conducting elections consistently reduce capital expenditure in the months before polling day [5]. They shed no light on why the most fiscally stressed Indian state, Punjab, with a debt-to-GSDP ratio of 48.1% and %40 of its revenues consumed by interest payments alone, arrived at that position through an unbroken cycle of competitive power subsidies begun in 1997 and ratcheted upward by every subsequent government regardless of party [6]. And they offer nothing to explain why the most expensive election in global history, the 2024 Indian general election with an estimated total expenditure of Rs 1.35 lakh crore, was conducted through a financing architecture in which 35% of party funds lacked identifiable sources, and a single anonymous donation scheme channelled Rs 12,979 crore before a unanimous Supreme Court struck it down as unconstitutional [7].

This article argues that India’s development paradox is not accidental but structural, generated by a political economy in which the logic of electoral competition consistently privileges short-run consumptive transfers over long-run public-goods provision. Three analytical channels drive this logic. The first is the legitimacy architecture of electronic voting machines: An unresolved credibility gap that diverts democratic energy toward contestation of electoral outcomes and enables losing parties to substitute legal challenge for programmatic accountability.

The second is the freebie spiral: The multi-decade escalation of unconditional pre-election transfers that has crowded out capital expenditure, compressed merit-good budgets, and accumulated subnational debt at a pace the Reserve Bank of India has repeatedly flagged as a systemic fiscal risk. The third is the money-election nexus: The integration of illegal vote-buying, informal cash distribution, and formally institutionalised anonymous corporate donations into the structure of electoral competition, producing a political class whose incentives are oriented toward rent extraction and short-term constituency management rather than developmental governance.

None of these channels is entirely new to the literature. The EVM controversy has a technical record going back to the landmark 2010 ACM paper by Wolchok et al. [1]. The economics of Indian populism has been examined by Devi [8], Mukhi et al. [9], and the RBI’s State Finances series. The money-politics nexus has been documented by the Association for Democratic Reforms across a sustained research programme [7]. What is new in this article is the integration of these three channels into a single analytical framework that explains, through the interaction effects rather than any channel in isolation, why the development equilibrium is stable rather than self-correcting.

Fig. 1 presents this framework schematically.

Arrows indicate causal pathways; double arrows indicate reinforcing feedback loops between channels. Constructed by the author on the basis of the theoretical and empirical literature reviewed in Sections 2 through 6.

The remainder of the paper proceeds as follows. Section surveys the relevant literature across the three channels and positions this article’s contribution. Section examines the EVM architecture, the security vulnerability record, and the 2024 judicial rulings. Section analyses the freebie spiral from definitional grounds through its fiscal anatomy and development trade-offs, drawing on RBI, PRS Legislative Research, and state budget data.

Section investigates informal vote-buying and the formal corporate donation architecture. Section traces the development consequences of the composite political economy. Section develops a five-pillar reform agenda. Section addresses limitations and competing interpretations. Section concludes.

The institutional design of ballot technology carries consequences beyond administrative efficiency: it shapes the perceived legitimacy of outcomes, and that perception is itself a constitutive component of democratic stability [10]. The distinction between actual integrity and perceived integrity matters above all in polities where electoral margins are narrow and losing parties face strong incentives to contest results. Norris [11] establishes that the credibility of electoral technology depends not only on its technical robustness but on the transparency of the institutions that certify it. A voting system that is technically secure but institutionally opaque can generate legitimacy deficits as damaging to democratic practice as outright manipulation.

India’s EVM architecture has been scrutinised in this light since its nationwide adoption. The most technically rigorous published challenge remains the study by Wolchok et al. [1], presented at the ACM conference on computer and communications security, which obtained a physical Indian EVM and demonstrated two hardware-level attack vectors: replacement of display components with look-alike modules capable of storing and transmitting vote data, and extraction of stored vote counts using portable external devices. The Election Commission of India disputed the study’s conclusions, arguing that the machine in question may not represent current-generation hardware and that multi-layer custody procedures make such attacks infeasible at scale. Desai and Lee [12] review subsequent peer commentary and conclude that while the specific attack vectors identified by Wolchok et al. [1] may be less practicable with newer hardware, the core institutional concern persists: The ECI’s posture of asserting “infallibility” forecloses the adversarial open testing that is universally recognised in election security research as the only credible basis for assurance claims [13]. The social construction of technology analysis by Herstatt and Herstatt [14] traces how the EVM’s legitimacy became a contested symbolic resource in Indian politics, with the controversy’s persistence driven less by accumulating technical evidence of manipulation and more by institutional opacity and the asymmetric incentives of electoral losers.

The theoretical foundation for the study of electorally motivated public expenditure runs from Nordhaus [15]'s political business cycle model through Rogoff [16]'s signalling framework for electoral budget cycles and Persson and Tabellini [17] formal treatment of political economics in developing democracies. The India-specific literature on electoral cycles in public expenditure has grown substantially since the Supreme Court’s 2022 hearing on freebies and the RBI’s June 2022 state finance risk report [18]. Khemani [5], using panel data from 14 Indian states from 1960 to 1994, finds a statistically significant increase in expenditure on visible goods in the pre-election period. Dash and Mukherjee [19], in a NIPFP study of political competition and HDI across Indian states, establish that politically competitive governments perform better on HDI outcomes, but that this positive effect is conditioned on the type of competition: Systems where parties compete primarily on public goods provision outperform those where competition is primarily clientelistic. Devi [8] documents in detail for Delhi that welfare subsidy spending grew by 609% between FY2014–15 and FY2024–25 while capital expenditure contracted from 22.7% to 16.4% of the total budget. The SSRN working paper of Mukhi et al. [9] employs linear regression on seven-state panel data from the RBI to establish that social sector commitments, outstanding liabilities, and subsidies together produce statistically significant negative impacts on gross fiscal deficit, providing econometric confirmation of the fiscal sustainability concern. Kumar [20], writing in the International Journal of Management Research and Review Studies, uses a qualitative-comparative analysis of state election manifestos from 2019 to 2024 to identify 47 distinct categories of freebie promise across 19 elections, finding that the average fiscal cost per election has risen from Rs 3,200 crore in 2019 to Rs 8,700 crore in 2024.

The theoretical literature on vote buying in low-income democracies distinguishes persuasion models, in which payments shift sincere preferences, from turnout-buying models, in which payments mobilise already-committed partisans [21], [22]. Vaishnav [23] provides the most comprehensive book-length treatment of the money-crime-politics nexus in India, documenting through a combination of court records, candidate affidavit analysis, and survey data that voters in low-income constituencies systematically prefer candidates who can demonstrate both criminal capacity and economic resources, because these attributes credibly signal the ability to deliver patronage in a setting where formal institutional guarantees are weak.

The most rigorous econometric evidence on direct vote buying is Mitra and Mitra [24], published in Economic Development and Cultural Change, which identifies a statistically significant consumption spike in election-proximate periods concentrated in swing constituencies, consistent with strategic cash distribution rather than generalised income effects. Goyal [25], writing in the Journal of Politics, finds that voters in constituencies receiving programmatic investments in rural roads between elections are more likely to hold incumbents accountable for development outcomes, but that this accountability mechanism is crowded out by cash transfers in the pre-election period. The corporate finance dimension was brought to the fore by the supreme court’s February 2024 ruling in association for democratic reforms v. Union of India, which struck down the Electoral Bond Scheme; subsequent investigative analysis by Poola and Anna John [26], published in humanities and social sciences communications, documents four distinct quid-pro-quo modalities embedded in the scheme using a combination of company registration data, enforcement agency records, and contract award databases.

The foundational empirical work on electoral politics and public goods provision in India is Banerjee and Somanathan [27], which establishes using 1970–1980 data that the allocation of public goods across constituencies responds to ethnic fragmentation and electoral competition in ways that systematically disadvantage non-aligned voters. Keefer and Khemani [28] introduce the concept of credibility in electoral promises, arguing that where parties cannot credibly pre-commit to programmatic policies, they default to targeted transfers whose delivery can be individually monitored, producing a structural bias against broad public goods in competitive but weakly institutionalised democracies. The more recent study by Jensenius and Chhibber [29], published in Comparative Political Studies, exploits a natural experiment from Indian electoral redistricting to show that legislator identity affects the allocation of public spending in constituency-specific ways that compound historical disadvantage. Heath et al. [30], in World Development, document that credit attribution failures in India’s multi-level democracy systematically weaken voter capacity to reward health investment at the state level, creating a free-rider problem in health spending across the federal system. The development-economics context is provided by UNDP [2], whose HDI data show India’s ranking has moved from 132nd in 2019 to 134th in 2023 (a ranking shift reflecting expansion of the reference set rather than absolute deterioration), and by the ASER [4] learning poverty data. The IMF’s observation, cited in PRS [31], that most of the benefits from agricultural power subsidies accrue to higher-income farming households, directly challenges the distributional justification for the largest single category of Indian state freebie spending.

India’s EVM system, developed by Bharat electronics limited and electronics corporation of India Limited, two government-owned manufacturers, consists of a Ballot Unit accessed by the voter and a Control Unit operated by polling staff, joined by a five-metre cable. The voter-verifiable paper audit trail, mandated by the Supreme Court in 2013 and operationalised from 2014, adds a third component that prints a paper slip visible to the voter for seven seconds, providing independent confirmation of the recorded choice. The ECI’s canonical security claim rests on three technical properties: One-time programmability of embedded microcontrollers, meaning firmware is burned during manufacture and cannot be modified by software subsequently; the absence of any network, wireless, or internet connectivity; and a physical custody chain involving multi-party sealing, randomised constituency allocation, and first-level checking with candidate agents present.

The ECI’s official position, publicly asserted since at least 2009, is that EVMs are “infallible.” This claim is problematic not primarily because it is false but because it is unverifiable under the current institutional architecture. Cryptographic and computer security research has consistently held since the 1990s that security claims for any system are only credible when the system has been subjected to rigorous adversarial independent testing. The ECI’s long-standing reluctance to permit such testing, and the limited scope of testing when eventually permitted, constitutes an institutional transparency deficit that generates legitimacy damage quite independently of the technical reality [11].

The most consequential published technical assessment of Indian EVMs is the paper by Wolchok et al. [1] presented at the 17th ACM conference on computer and communications security. The study team, which included J. Alex Halderman of the University of Michigan, a leading global authority on election security, obtained a physical Indian EVM from an anonymous source and subjected it to hardware and procedural analysis. Two attack classes were demonstrated. In the first, the display board of the Ballot Unit was replaced with a look-alike component that stored vote data and transmitted it by radio to an external device; this attack required only that an insider have physical access to the machine for a few minutes during the custody chain. In the second, a “clip-on” memory-reading device was attached to the external surface of the Control Unit to extract and potentially modify the stored vote count; this attack was designed to exploit the EVM’s use of a serial communication protocol without cryptographic authentication.

The study concluded:

“We conclude that in spite of the machines’ simplicity and minimal software trusted computing base, they are vulnerable to serious attacks that can alter election results and violate the secrecy of the ballot [1].” The ECI’s response focused on the provenance of the tested machine and on the procedural safeguards, arguing that the multi-party sealing process makes hardware substitution detectable in practice. Independent commentary by Desai and Lee [12] acknowledges that the ECI’s procedural objections carry weight but notes that the test of any security architecture is not whether attacks are theoretically detectable but whether detection would reliably occur given actual enforcement capacity across 1.4 million EVMs deployed across half a million polling booths by a temporary workforce of millions of polling officers. The gap between theoretical safeguards and practical enforcement capacity is itself a structural vulnerability. Wolchok et al. [1] subsequently identified additional design concerns related to the absence of cryptographic integrity checks on inter-unit communication, concerns that apply regardless of the custody question.

The February 2024 Jantar Mantar protest in New Delhi, organised by a coalition of opposition parties and civil society organisations under the banner of the #EVM_Ban campaign, drew thousands of participants demanding a return to paper ballots ahead of the April general election [32]. The protest was preceded by a wave of social media content and a viral video in which a civil engineer claimed to demonstrate that EVM microcontrollers could be reprogrammed using commercially available tools, a claim that was never independently verified but circulated widely nonetheless.

What is analytically significant about this episode is not the technical accuracy of the claims made but the scale and speed of their uptake, which reflects a reservoir of institutional distrust that the ECI’s “infallibility” posture has accumulated over two decades. Bawa et al. [32] document through survey data that EVM-related scepticism is statistically significantly higher among voters from marginalised communities, not because marginalised voters possess technical knowledge of machine design but because their historical experience of institutional betrayal makes them more susceptible to narratives of systemic manipulation. The political economy implication is that EVM distrust functions as a mobilisational resource for parties whose primary appeal is to these communities, creating a durable incentive to sustain rather than resolve the controversy.

Two Supreme Court judgments in 2024 defined the current legal landscape. In April, a division bench comprising Justices Sanjiv Khanna and Dipankar Datta dismissed petitions by the Association for democratic reforms and the National Congress Party seeking 100% VVPAT slip counting, a return to paper ballots, or voter self-verification of VVPAT output. The Bench conducted a technically detailed hearing, requiring the ECI to explain microcontroller one-time programmability, machine sealing procedures, and the statistical basis for the existing VVPAT sampling rate. The Court directed two modifications: Symbol Loading Units used in VVPAT machines are to be sealed and stored for 45 days post-result, and EVM mock polls are to be available at candidate request. The ECI submitted that 41,629 instances of random verification had been conducted with no recorded discrepancy attributable to machine malfunction, and that across the five-phase 2024 general election, 38,156 VVPAT paper slips were matched against EVM counts without anomaly. In November 2024, the Court dismissed a PIL seeking a return to paper ballots, reinforcing its April position but adding an observation that EVM-related challenges “consistently emanate from electoral losers” and that the judiciary’s role is to adjudicate legal questions rather than serve as a mechanism for relitigating electoral outcomes. The November ruling implicitly acknowledges, without fully addressing, the asymmetry identified in the security literature: the burden of proof for demonstrating manipulation is structurally allocated to challengers who have no access to the machines, while the institution with access controls its own accountability.

Events above the baseline represent institutional and legislative developments; events below represent civil-society challenges, security studies, and judicial interventions. Lower panel: Trajectory of composite public trust in EVMs and the Election Commission of India based on synthesised survey evidence from Lokniti-CSDS, ADR, and published opinion research. Note that trust scores represent relative trajectories constructed from published survey summaries and should be interpreted directionally rather than cardinally (Fig. 2) [1], [14], [32].

India’s EVMs appear to be more technically resistant to manipulation than the public controversy implies, but the controversy itself causes measurable democratic damage by suppressing voter confidence among marginalised communities and enabling losing parties to delegitimise electoral outcomes without bearing the costs of producing evidence. The Supreme Court’s April 2024 ruling reduces but does not eliminate this legitimacy deficit because it upholds current procedures without addressing the systemic transparency gap identified in the independent security literature. The resolution requires not adjudication of the technical debate but a transition toward independently verifiable cryptographic audit standards that would render the debate empirically moot.

Definitional clarity: The analytical utility of the freebie category depends on distinguishing unconditional electoral redistribution from merit-good provision and from conditional social protection. The RBI defines freebies as “welfare measures provided free of charge by the government, as distinguished from merit goods whose provision generates positive externalities justifying subsidy on social efficiency grounds” [18]. The Supreme Court’s 2022 observations in Ashwini Upadhyay v. Union of India identified “appeasing policy” as a fiscal category distinct from developmental expenditure without providing a legally binding definition, and the Court’s January 2026 ruling established a clearer distinction between mass distribution of state funds to individuals and investment in public welfare.

For the purposes of this article, a freebie is an unconditional or weakly conditioned transfer of goods, cash, or services to a broad voter group, announced proximate to an electoral event, without a published fiscal sustainability assessment or integration into a medium-term developmental framework. By this definition, the following categories are analysed: Free electricity beyond a minimal social tariff, monthly cash allowances for demographic categories, laptop and durable-goods distributions, blanket farm loan waivers, and waived property tax or water charges. The National Food Security Act’s subsidised grain provision, PM-KISAN income support, and Ayushman Bharat health insurance are explicitly excluded as they operate under distinct institutional frameworks.

Fiscal magnitude: The national picture: The scale of freebie-related fiscal commitments has grown faster than state revenues across the political spectrum. PRS legislative research’s state of state finances 2024–25 report documents that 24 states spent Rs 3,18,815 crore on subsidies in FY2023–24, equivalent to 9% of their aggregate revenue receipts. This figure has risen at a compound annual growth rate of approximately 14% since FY2018–19, compared to a CAGR of 8% for capital expenditure across the same states over the same period. The Economic Survey 2025–26 estimates that unconditional cash transfers and freebie schemes implemented across Indian states will cost approximately Rs 1.7 lakh crore in FY2026 alone, a figure equivalent to the entire central government allocation for capital formation in railways and urban infrastructure combined [3].

Eleven states estimated a revenue deficit in 2024–25 at the budget stage, meaning they must borrow to finance current expenditure, with no asset creation. The combined gross fiscal deficit of states rose from 2.6% of GDP in FY2022 to 3.2% in FY2025, with outstanding liabilities reaching 28.1% of GDP. Within the sub-national system, power subsidies constitute the dominant category of freebie expenditure: PRS documents that 19 states collectively allocated 53% of all subsidy expenditure to power subsidies in FY2023–24, with Punjab at 21% of revenue receipts, Tamil Nadu at 14%, and Andhra Pradesh at 12% leading the distribution. Table 1 provides a state-level cross-section of the key fiscal stress indicators.

The Punjab case: A cautionary arithmetic: Punjab represents the most advanced stage of the freebie spiral currently observable within the Indian federal system. The state’s total debt stood at Rs 3,82,935 crore as of March 2025, with projections from the state government’s own budget placing it at Rs 4,17,136 crore by the end of FY2026. Nearly 40% of the state’s revenue is consumed by interest payments alone.

Development projects have suffered directly: The Centre cancelled road works worth more than Rs 800 crore under the Pradhan Mantri Gram Sadak Yojana-III, citing state delays in executing tenders attributable to fiscal incapacity.

The direct trigger for the latest acceleration was the Aam Aadmi Party government’s 2022 introduction of 300 units of free electricity per month to domestic consumers, a scheme with an estimated annual cost of Rs 8,785 crore projected for FY2025, up from Rs 5,739 crore in FY2022–23. Combined with the pre-existing free power scheme for agricultural users, which costs Rs 8,809 crore per year and has driven the number of tubewell connections from 2.8 lakh in the 1980s to 14.5 lakh by 2023, Punjab’s total power subsidy bill is projected to touch Rs 20,000 crore in FY2025. The environmental consequence is severe:

14.5 lakh tubewells pumping on subsidised electricity extract an estimated 4,385 billion litres of water per week, driving a groundwater depletion crisis that threatens Punjab’s long-run agricultural productivity independently of the fiscal question.

The political-economy logic is documented by economist Ghuman [33], whose study understanding economic acceleration and deceleration of Punjab traces the debt accumulation to competitive populism across party lines: The free electricity cycle was initiated in 1997 by Chief Minister Rajinder Kaur Bhattal, expanded to all farmers by Parkash Singh Badal regardless of land size, extended to industries by the Akali-BJP coalition with attendant tax concessions, accelerated by the Congress government’s Rs 20,000 crore annual borrowing during 2017–22, and ratcheted further by the AAP government’s domestic free power scheme. No government of any ideological alignment has moved to dismantle the cycle, because the electoral cost of doing so would be immediate while the fiscal benefit would accrue over years.

The crowding-out mechanism: Fig. 3 illustrates the fiscal anatomy of the freebie spiral across the major states from FY2018 to FY2025. Panel B shows the systematic divergence between freebie and subsidy allocations, which have grown from 7.9 to 14.8% of total state expenditure, and capital expenditure, which has contracted from 15.1% to 13.4% over the same horizon. Panel C makes Punjab’s trajectory visible in its full severity.

Panel A: Subsidy expenditure as a percentage of revenue receipts by state in FY2023–24 (CAG actuals), with 21 of Punjab’s revenue receipts consumed by subsidies against a national aggregate of 9%. Panel B: Time series of freebie and subsidy allocations versus capital expenditure as shares of total state expenditure from FY2018 to FY2025, showing systematic divergence. Panel C: Punjab’s total outstanding debt from 2015 to 2026, with the AAP-period free electricity scheme marked; the state’s own budget projects debt reaching Rs 4.17 lakh crore by FY2026 (Fig. 3) [6].

The crowding-out mechanism operates through three pathways. Directly, when current transfers pre-commit revenue streams, capital budgets become the residual claimant and are cut first because they lack a vocal constituency of immediate beneficiaries. Indirectly, growing subnational borrowing raises the aggregate cost of sovereign debt, crowding out private investment through interest rate effects. Through the fiscal-political channel, the accumulation of off-budget liabilities, particularly power utility debt not reflected in official deficit numbers, conceals the true fiscal cost of freebie programmes until a crisis point, by which time the reforms required are politically even more costly than they would have been earlier.

In FY2023–24, 24 Indian states spent Rs 3,18,815 crore on subsidies, 9% of aggregate revenue receipts. This amount represents a 14% CAGR since FY2019, more than double the growth rate of capital expenditure. Punjab’s debt-to-GSDP ratio of 48.1%, driven by competitive power subsidies initiated in 1997 and never reversed by any subsequent government, represents a near-terminal case of what Narayan [6] calls a “vicious cycle of utility debt, deteriorating service quality, and increased contingent liabilities.” The RBI’s 2022 and 2024 State Finances reports [18] have used unusually direct language in identifying rising subsidy burdens as a systemic fiscal risk. The Economic Survey 2025–26 estimates total freebie costs at Rs 1.7 lakh crore in FY2026 [3].

The Sri Lanka comparison and the Boundary condition: The relevance of the Sri Lankan economic collapse of 2022 to India’s freebie debate has been noted in both the popular press and the academic literature [8]. Sri Lanka’s trajectory, tax cuts combined with accelerating welfare entitlements, rapidly rising dollar-denominated debt, and a foreign exchange crisis triggered by the pandemic shock, is not mechanically replicable at the Indian federal level because India’s sovereign debt is primarily rupee-denominated and the central government retains substantially stronger fiscal buffers.

The comparison is, however, directly applicable at the state level for the most stressed Indian states, where the dynamics of revenue deficit, power sector debt, and rising interest payments as a share of revenue create conditions qualitatively comparable to early-stage Sri Lankan fiscal stress. The fifteenth finance commission specifically highlighted how delayed subsidy payments to power utilities create a vicious cycle that reproduces Sri Lanka-like dynamics at subnational scale [6].

The scale of electoral expenditure: The 2024 Indian general election was the most expensive electoral contest in human history measured on a per-voter basis. The Centre for Media Studies, which has tracked Indian election expenditure for 35 years, estimated total spending at Rs 1.35 lakh crore, revised upward from an initial estimate of Rs 1.2 lakh crore after the electoral bond disclosures emerged.

This figure encompasses expenditure by political parties, individual candidates, government advertising, media placement, and unaccounted cash whose volume can only be inferred. The gap between declared and actual expenditure is enormous: Official party expenditure reports submitted to the ECI showed total declared spending of Rs 3,352 crore across all 32 parties that submitted accounts, with the BJP accounting for Rs 1,493 crore or 44.6% of declared total. Against an estimated actual total of Rs 1.35 lakh crore, the declared figure represents barely 2.5% of estimated actual spending, a ratio that quantifies the systemic opacity of Indian electoral finance more powerfully than any normative critique.

Across the six general elections from 1999 to 2024, estimated total election expenditure has risen from approximately Rs 4,000 crore to Rs 1,35,000 crore, a 34-fold increase that substantially exceeds nominal GDP growth over the same period. The ECI’s own figures on election-time seizures of cash, liquor, drugs, and gold, which represent actual confiscation of intended vote-buying material, crossed Rs 9,246 crore in 2024, compared to Rs 180 crore in 2004, a 51-fold increase.

These seizures, while large in absolute terms, almost certainly represent a small fraction of actual distribution, given the constraints on ECI enforcement capacity across half a million polling booths staffed primarily by temporary state government employees embedded in local political networks. Table 2 summarises the electoral finance landscape across the 2024 cycle.

The financial architecture of Indian electoral competition: Scale, opacity, enforcement, and wealth (1999

Panel A: Estimated total election expenditure versus officially declared party expenditure across general elections, 1999–2024 (log scale), illustrating the scale of the transparency gap. Panel B: Distribution of electoral bond proceeds by party, FY2017–2024, prior to the scheme’s striking down by the Supreme Court on 15 February 2024. Panel C: ECI enforcement seizures of cash, alcohol, drugs, and gold across general elections on a log scale, showing 51-fold growth from 2004 to 2024. Panel D: Declared asset comparison between elected legislators and average citizens, illustrating the systematic concentration of economic power among elected representatives.

Direct vote buying: Evidence from household data: The most rigorous published evidence on direct cash distribution to voters is the study by Mitra and Mitra [24], which uses a difference-in-differences design combining National Sample Survey Office household consumption data with state assembly election timing across the period 2004 to 2012. The study identifies a statistically significant positive effect of election proximity on reported household consumption, concentrated in swing constituencies rather than safe seats or uniformly poor districts. The spatial pattern, which distinguishes strategically targeted swing voters from committed partisans and geographically disadvantaged but non-competitive areas, is consistent with the theoretical prediction of the strategic vote-buying models developed by Stokes [21] and Nichter [22]. The fact that the consumption spikes identified by Mitra and Mitra [24] exceed concurrent wage bill changes for the same districts rules out labour-market explanations and strongly implicates pre-election cash transfers as the operative mechanism.

The pattern documented in academic research is consistent with what ECI enforcement data reveal. In Assembly elections held in 2023, Rs 2,014 crore was seized by ECI flying squads and surveillance teams, compared to Rs 239 crore in corresponding elections in the prior cycle, a 742% increase in a single electoral cycle that reflects both the escalating scale of distribution and improved enforcement capacity. The ECI deployed 6,398 district nodal officers and 59,000 flying squads and static surveillance teams across the 2024 general election, an enforcement apparatus of formidable scale that nonetheless intercepted at most a single-digit percentage of total cash in circulation given the Rs 9,246 crore seized against an estimated Rs 20,000 crore or more in annual unaccounted cash entering the electoral system according to ADR’s estimate. The electoral bond scheme: Architecture and exposure: The electoral bond scheme, introduced by the Government of India in January 2017 through amendments to the Finance Act, the Representation of the People Act, the Reserve Bank of India Act, the Companies Act, and the Income Tax Act, was the most significant institutional transformation in Indian political finance since the introduction of comprehensive disclosure requirements in 2004. Bonds were sold exclusively through the State Bank of India in denominations from Rs 1,000 to Rs 1 crore. Any company registered in India for at least three years could purchase them without disclosure, and they were redeemable only by registered political parties within fifteen days. No donor identity needed to be disclosed to the ECI, to the public, or to other parties, and the corporate donation cap of 7.5% of three-year average net profit was removed, enabling unlimited corporate donations. Between January 2017 and March 2024, bonds worth Rs 12,979 crore were sold and redeemed. Approximately 57%, or Rs 8,251 crore, went to the BJP; Rs 1,610 crore to the Trinamool Congress; Rs 1,421 crore to the Indian National Congress; and Rs 1,215 crore to the Bharat Rashtra Samithi, with the remaining Rs 482 crore distributed across 14 other parties. The SBI’s disclosure to the Supreme Court following the February 2024 ruling enabled investigative analysis by Poola and Anna John [26], published in Humanities and Social Sciences Communications, which documents four distinct corruption modalities. Companies under active enforcement investigation by the CBI, ED, or income-tax authorities purchased bonds disproportionately in the months following the initiation of investigations. Newly established or small companies with revenues insufficient to explain their bond purchases donated large sums. Cross-party transfers among bonds suggest informal market-making. Companies receiving government contracts above threshold values showed statistically anomalous bond purchase timing relative to contract award dates.

The constitutional Bench’s unanimous ruling, authored by Chief Justice D.Y. Chandrachud with four concurring justices, held that the scheme violated voters’ right to information under Article 19(1)(a) of the Constitution, that the removal of corporate donation caps without adequate justification enabled corruption, and that the combination of unlimited anonymous corporate donations and the government’s unique access to SBI transaction data created a “structural asymmetry of information” that was “manifestly arbitrary”. Criminalisation of politics and the Vaishnav [23] Thesis: The money-politics nexus in India extends beyond electoral finance to the systematic entry of individuals with criminal backgrounds into elected office. Vaishnav [23] argues that this is not simply a failure of voter discrimination but a rational equilibrium: voters in low-income constituencies with weak state delivery of public services rationally prefer candidates who can demonstrate the capacity to deliver patronage through a combination of economic resources and coercive capacity, because formal institutional channels for public goods delivery are unreliable. ADR data from the 2024 election show that 46% of winning candidates in the Lok Sabha declared criminal cases against themselves, and 29% declared serious criminal cases including charges of murder, kidnapping, or extortion. More recently, Khemka [34], writing in the European Journal of Political Economy, provides a formal model of why voters elect criminal politicians that grounds the outcome in an information asymmetry between voters and candidates combined with the credibility advantage that criminal records provide as signals of delivery capacity in weak-institution environments.

India’s 2024 general election cost an estimated Rs 1.35 lakh crore, making it the most expensive electoral contest in global history on a per-voter basis. Officially declared party spending represents 2.5% of this estimate. ECI enforcement seized Rs 9,246 crore in vote-buying material in 2024 alone. A formally institutionalised anonymous donation scheme channelled Rs 12,979 crore in opaque corporate money to political parties before being struck down as unconstitutional. The average elected Member of Parliament holds declared assets 27 times the wealth of the average urban Indian household. The combined formal and informal financial architecture creates systematic incentives for transactional governance and against programmatic public-goods delivery.

The HDI-GDP divergence: Structure of the paradox: India’s per capita GDP in purchasing-power-parity terms grew from USD 443 in 2000 to approximately USD 2,600 in 2023, a nearly six-fold increase that reflects genuine and substantial improvements in material living standards, poverty reduction, and structural economic transformation. India’s HDI over the same period rose from 0.496 to 0.644, an improvement of 30% in absolute score. The paradox is not that development has been absent but that the relationship between income growth and human development outcomes is dramatically weaker in India than in comparator economies.

Vietnam, with a similar per capita income, has an HDI of 0.726. Sri Lanka, before its fiscal crisis, was at 0.780. Indonesia is at 0.713. The systematic underperformance of India’s HDI relative to its income level reflects a public investment composition that consistently prioritises infrastructure and transfers over human capital formation [2].

Development consequences of the electoral political economy. Panel A: India’s HDI score trajectory from 2000 to 2023 plotted against per capita GDP growth, illustrating the systematic divergence between income expansion and human development improvement (Fig. 5). Panel B: comparative public spending on education and health as percentages of GDP across India and peer economies, with India’s NEP 2020 targets shown; India’s underinvestment relative to comparators is stark. Panel C: ASER rural learning outcome data from 2012 to 2022 showing the fraction of Class 5 students achieving grade-level reading and arithmetic skills; the flat trajectory across a decade of nominal educational spending increase documents learning poverty at scale [2–4].

Education: Learning poverty at scale: India’s education spending at 2.9% of GDP is not only below the 6% target of NEP 2020 but below the level India itself spent in the 1990s, before the competitive freebie spiral gained momentum. Brazil spends 6.5%, China 4.2%, and the OECD average is 5.1%.

The consequences are visible in ASER data. The ASER 2022 report [4], based on household surveys of 700,000 children across 616 rural districts, found that only 42.8% of Class 5 students could read a Class 2- level text, and only 25.9% could perform a simple three-digit division problem. These proportions are statistically indistinguishable from the corresponding figures for 2012, meaning that a decade of nominal increases in school enrolment, mid-day meal programmes, and educational infrastructure spending has produced essentially zero improvement in foundational learning outcomes in rural India [4]. The NITI Aayog has specifically identified durable-goods giveaways such as laptop distributions as consuming fiscal resources that would generate higher returns if deployed on teacher training and instructional quality improvement, a finding that directly connects the freebie category to the learning poverty outcome.

The sectoral mechanism is not difficult to trace. Teacher recruitment and training, instructional material development, and school quality monitoring all require recurrent discretionary budget allocations that can be quietly reduced without immediate visible consequence. Laptop distributions and free uniform schemes, on the other hand, generate visible credit for which a minister can take a photograph. Over a decade, that political preference for the photographable transfer over the invisible investment in quality produces exactly the ASER outcome we see.

Healthcare: Structural underfunding: India’s public health expenditure at 2.1% of GDP is the second-lowest in the G20, ahead only of Indonesia, and less than one-quarter of the OECD average of 8.8%. Brazil spends 9%, Thailand 4.3%, and China 5%. The consequences were most dramatically exposed by the COVID-19 Delta wave of April to June 2021, which overwhelmed India’s healthcare system not because India lacked the financial resources to have built adequate capacity but because successive governments had consistently deferred health infrastructure investment in favour of current transfers. The absence of adequate ICU beds, oxygen supply chains, and trained critical-care personnel in the world’s fifth-largest economy was a direct consequence of the political economy documented in this article [30].

The IMF, cited by PRS [31], observes that most of the consumption benefits from agricultural power subsidies accrue to higher-income farming households who own more land and use more electricity, meaning that the largest single category of Indian state freebie spending is both fiscally unsustainable and regressive in distributional terms, simultaneously failing the fiscal sustainability test and the distributional equity justification typically offered in its defence. Replacing agricultural power subsidies with direct income support targeted by landholding size, as the fifteenth finance commission recommended [6], would address both failures simultaneously but has not been implemented by any major state because the political coalition supporting free power includes both small farmers, who benefit modestly, and large agricultural interests and groundwater extraction industries, who benefit substantially and contribute disproportionately to party finances.

Infrastructure, private investment, and the fiscal multiplier: The macroeconomic consequence of capital expenditure compression extends beyond the direct underprovision of public goods. Public capital has well-documented complementarity with private investment in developing economies: Roads reduce logistics costs and enable private firms to invest in production capacity, irrigation reduces agricultural risk and incentivises crop intensification, and broadband infrastructure enables human capital accumulation and e-commerce. The fiscal crowding-out of public capital therefore generates a private investment multiplier effect in reverse. In a study of 17 major Indian states, find that infrastructure provision has a significant negative relationship with regional inequality, meaning that the contraction of capital expenditure in high-freebie states has inequality consequences beyond the direct fiscal channel. India’s logistics costs at 14% of GDP are nearly double the global average of 8%, in part because the road and port infrastructure that would reduce them has been underbuilt relative to what India’s income level would predict, a gap partly attributable to the systematic compression of capital budgets.

India’s political economy of electoral competition generates a composite development loss through four interacting mechanisms. Consumptive transfers crowd out capital formation in public infrastructure, compressing the physical capital stock. Fiscal stress suppresses merit-good spending on education and health, compressing the human capital stock. Corporate capture through formal and informal electoral finance channels distorts regulatory and contracting decisions away from productivity-enhancing allocations. The legitimacy deficit of the EVM system diverts democratic energy toward outcome contestation rather than developmental accountability. The joint effect is India’s stable low-development equilibrium: HDI rank 134, learning poverty in rural schools near 60%, healthcare spending at 2.1% of GDP, and infrastructure logistics costs double the global average.

The analytical framework developed in this article points toward a five-pillar reform agenda, each pillar targeted at one or more of the transmission mechanisms identified.

Pillar I: Electoral technology transparency: The most technically tractable reform to the EVM legitimacy deficit is the adoption of end-to-end verifiable voting, a cryptographic framework in which the integrity of ballot aggregation can be mathematically verified by any observer with a standard computer, without requiring trust in either the EVM hardware or the custody chain managed by the institution being audited [13]. Estonia, Switzerland, and Norway have deployed such systems at national scale. A staged transition in which each EVM generates a cryptographic receipt independently verifiable by candidates and observers would preserve the speed and administrative advantages of the current system while providing an assurance standard the VVPAT sampling approach cannot match.

Pending that transition, the expansion of VVPAT sampling from the current 0.1% to 0.3% of booths to a statistically powered sample designed by independent statisticians to detect manipulation at a specified confidence level, with the sample size calculation published before each election, would substantially improve both actual security and the credibility of the ECI’s assurance claim.

Pillar II: Manifesto costing and fiscal accountability: The most direct point of intervention in the freebie spiral is the pre-election period when parties announce commitments without publishing credible cost estimates. A mandatory manifesto costing regime, in which parties must submit fiscal cost assessments of all commitments to an independent technical panel, comprising the 16th Finance Commission, the CAG, and NITI Aayog, within 48 hours of publication, and must receive certification that the commitments are consistent with the state’s medium-term fiscal framework before they can be publicised, would insert an accountability moment between electoral promise and budget reality. The European Union’s Stability and Growth Pact provides one template; the Netherlands’ CPB Bureau for Economic Policy Analysis provides a more direct model, in which all major parties submit manifestos for costing and the results are published jointly before the election. Pillar III: Independent fiscal councils: The existing Fiscal Responsibility and Budget Management Act architecture has been repeatedly relaxed, amended, or exempted by successive central and state governments, demonstrating that statutory rules without enforcement independence are insufficient. An independent fiscal council with statutory powers to publish binding assessments of state government budgets, similar to the UK office for budget responsibility or the US congressional budget office, would provide the external discipline that electoral cycles consistently erode. The 16th Finance Commission could be tasked with designing the architecture and certifying fiscal council independence standards.

Pillar IV: Electoral finance architecture: The post-electoral bond landscape creates an opportunity to build a political finance system that is simultaneously transparent and practically functional. Three design elements are essential. All corporate and individual donations above an aggregate annual threshold of Rs 20,000 per party should be disclosed within 30 days to a public real-time registry administered by the ECI. Corporate donation limits should be restored at a level tied to verified average net profit over five years to prevent shell-company vehicles. A system of conditional public funding for registered parties, based on vote share in the previous election and conditioned on annual audit compliance and internal democratic governance certification, would reduce party dependence on large corporate donors by providing a predictable funding base, following the German Parteienfinanzierung model.

Pillar V: Demand-side democratic accountability: Supply-side institutional reforms will not alone resolve a political economy in which voters rationally prefer immediate transfers to long-run public goods when they cannot credibly monitor whether productive public investment will occur or benefit them. Three demand-side instruments have demonstrated effectiveness in Indian contexts. Social audits of public expenditure, in which community groups with statutory access to original vouchers and payment records scrutinise scheme implementation, have measurably reduced MGNREGS leakage in Andhra Pradesh [35].

Constituency-level development scorecards, tracking school learning outcomes, infant mortality, road conditions, and water access by parliamentary constituency, published by an independent agency, would enable voters to compare their representative’s promise-delivery record across election cycles, following the Brazilian Bolsa Familia evaluation model. Strengthened candidate criminal background disclosure, building on the ADR litigation since 2018, with sanctions for non-disclosure rather than merely formal requirements, would improve the signal quality of candidate information available to voters.

I. Pillar I (electoral technology): End-to-end verifiable cryptographic voting; statistically powered independent VVPAT audit protocols.

II. Pillar II (manifesto costing): Mandatory pre-election fiscal costing by independent technical panel; certification requirement before publication.

III. Pillar III (fiscal councils): Independent statutory councils for subnational budget assessment; binding reporting obligations.

IV. Pillar IV (electoral finance): Real-time donation disclosure registry; restored corporate donation limits; conditional state party funding.

V. Pillar V (demand-side accountability): Statutory social audits; constituency development scorecards; enhanced criminal disclosure sanctions.