Liberalisation and Economic Reforms

India did not choose liberalisation; it ran out of the money required to postpone it. That origin explains almost everything that followed — why the reforms were deep in some places and untouched in others, and why no party has ever won an election on them. The 1991 programme is at once the most consequential economic decision of independent India and the least democratically authorised, and that tension is the subject of this article.

The economy that 1991 inherited

The architecture of control

The pre-reform regime is best understood not as socialism but as a system of permissions. The state did not own most of the economy; it decided who could produce what, how much, where, with which technology, financed by whom, and at what price.

• The Industries (Development and Regulation) Act, 1951 required a licence to establish a new undertaking, to expand capacity, to change the product mix and to shift location. Because licences were finite and applicants were not, the permission itself became a tradable asset — the licence-permit-quota raj, enforced by the inspector raj.
• The Monopolies and Restrictive Trade Practices Act, 1969 subjected firms above an asset threshold to separate clearance for expansion, merger and new undertakings. The perverse result was that the law penalised size rather than conduct: an efficient firm that grew hit the ceiling, a cartel of small inefficient firms did not.
• The Foreign Exchange Regulation Act, 1973 capped foreign equity at 40% and required companies to dilute or leave — IBM and Coca-Cola left in 1977.
• Import licensing and canalisation meant that most imports required a licence tied to an “actual user” condition, and that a list of bulk commodities could be imported only through state trading canalising agencies. Tariffs on top reached peak rates above 300%, with a simple average well over 100% in the late 1980s.
• The small-scale industry reservation list — at its height around 870 products, including garments, toys, leather goods and light engineering — legally barred large firms from making precisely the labour-intensive goods that every East Asian economy used to industrialise.
• Administered interest rates and directed credit completed the system. After bank nationalisation in 1969 and 1980, the Reserve Bank set deposit and lending rates, the statutory liquidity ratio and cash reserve ratio pre-empted a rising share of deposits for government borrowing, and priority-sector lending targets directed the rest. Capital was allocated by instruction, not by return.

The “Hindu rate of growth” and what was actually meant by it

• The phrase belongs to the economist Raj Krishna, who used it to describe the roughly 3.5% average growth of the 1950s to the 1970s — with population rising about 2.2%, leaving per capita growth near 1%: living standards doubling in a lifetime rather than a decade.
• The label was an observation about persistence, not about religion. His point was that the rate was stable across very different governments, monsoons and plans, which pointed to a structural rather than a cyclical cause.
• What made the number embarrassing was regional comparison. South Korea, Taiwan and Singapore grew at two to three times that rate from comparable starting points, by exporting the labour-intensive manufactures India had reserved for small firms.

The caged tiger

• India was described as a “caged tiger” — a potential held down not by poverty of resources but by the populist policy cage built around it, so that the country waited for a crisis rather than choosing reform. It was an economy whose potential was visible in its scientific establishment, entrepreneurial classes and market size, and whose performance was held down by the policy cage around it.
• Amartya Sen’s response to the metaphor is the necessary corrective: once the cage was opened the sprint was real but not what had been predicted, because a tiger that has been undernourished, unschooled and untreated does not run merely because the door is open.

The 1980s: reform before the reform

• Indira Gandhi’s return in 1980 began the loosening. Her government delicensed around twenty industries, introduced broadbanding — letting a licensed firm vary its product mix within a broad category without fresh clearance — relaxed MRTP asset limits, and raised external commercial borrowing.
• Rajiv Gandhi after 1985 went further. The 1985 budget, presented by V.P. Singh, cut personal and corporate tax rates and simplified excise through MODVAT; twenty-five more industry groups were delicensed; and the technology missions on drinking water, immunisation, literacy, telecommunications, oilseeds and dairy created a template of goal-directed public action outside the plan machinery.
• The Long-Term Fiscal Policy of 1985 was the first attempt to state a multi-year fiscal trajectory and signal tax rates in advance rather than by annual surprise.
• Growth in the 1980s averaged about 5.6%, decisively above the earlier trend and before any of the 1991 measures.

The Rodrik–Subramanian argument and the case against it

• Dani Rodrik and Arvind Subramanian argued that India’s growth transition dates to around 1980, not 1991, and was triggered not by liberalisation but by an attitudinal shift.
• The counter-argument is decisive on one point. The 1980s expansion was financed by borrowing, not by earnings. External debt rose from about $20 billion in 1980-81 to $70 billion by 1990-91, the fiscal deficit widened, and the current account was covered increasingly by commercial borrowing and non-resident deposits rather than exports.
• The honest synthesis is therefore sequential rather than either-or. The 1980s produced the growth acceleration and the crisis that ended it; 1991 produced the structural change that made growth sustainable. A pro-business loosening without external competitiveness can raise output for a decade and still hit a payments wall.

“No power on earth can stop an idea whose time has come. I suggest to this august House that the emergence of India as a major economic power in the world happens to be one such idea. Let the whole world hear it loud and clear. India is now wide awake.” — Manmohan Singh, budget speech of 24 July 1991

The crisis of 1991

The arithmetic of insolvency

• The combined fiscal deficit of the Centre and states exceeded 10% of GDP, with the Centre’s own gross fiscal deficit above 8% of GDP in 1990-91.
• The current account deficit reached about 3% of GDP in 1990-91, roughly $9.7 billion, at a time when there was no non-debt capital inflow to finance it.
• Foreign exchange reserves fell to about $1.1 billion by mid-1991 — enough for roughly two weeks of imports.
• The credit-rating downgrade to below investment grade was the mechanism of collapse. Once India lost its rating, commercial borrowing dried up and non-resident Indian deposits began flowing out rather than in, converting a financing problem into a run.
• Gold was pledged. In May 1991 the Chandra Shekhar government had the State Bank of India dispatch 20 tonnes of confiscated gold to the Union Bank of Switzerland as a sale with repurchase option; in July 1991 the Reserve Bank shipped 46.91 tonnes to the Bank of England and the Bank of Japan, raising $405 million, redeemed between September and November 1991.

The external shocks

• The Gulf War of 1990-91 delivered a double blow. Oil prices roughly doubled, adding some $2 billion to the import bill in a year, while remittances from Kuwait and Iraq stopped and the government financed the airlift of over 170,000 Indians from the Gulf.
• The collapse of the USSR ended the rupee-rouble trade arrangement under which India had settled a substantial share of its trade in non-convertible rupees. — abuse of dominance, anti-competitive agreements, combinations above a threshold

The political vacuum

• The crisis peaked under the weakest governments India had had. V.P. Singh’s National Front fell in November 1990; Chandra Shekhar governed with 64 seats on Congress support and resigned in March 1991 without passing a full budget; the assassination of Rajiv Gandhi on 21 May 1991 occurred mid-election.
• P.V. Narasimha Rao took office in June 1991 leading a minority government, with Manmohan Singh as Finance Minister — a technocrat with no political base, which was itself an asset. The reforms were launched by the government with the least apparent authority to launch them.

The IMF programme and conditionality

• India drew on an IMF standby arrangement approved in October 1991 and took World Bank structural adjustment lending, with conditionality covering fiscal correction, trade liberalisation and financial-sector reform.
• The politically significant feature is that the conditionality was never publicly acknowledged as the driver. The government presented the measures as its own design, which preserved sovereignty in appearance and made the domestic sale possible — at the cost that the reforms were never argued for on their merits before the public.

Internal or external causation: a genuine dispute

Position	Core claim	Strongest evidence	Weakness
Externalist	The crisis was caused by shocks — Gulf War, USSR collapse, oil prices — and reform was imposed by the IMF	The timing is exact; conditionality is documented	Shocks explain the timing of the collapse, not the fragility that made a modest shock fatal
Internalist	The crisis was the terminal stage of the 1980s debt-financed growth model; shocks merely triggered it	Fiscal and debt deterioration is visible from 1985; other oil importers absorbed the same shock	Understates how narrow the actual escape window was
Ideational	An epistemic shift among policy elites predated the crisis and the crisis merely opened the window	Reform blueprints existed within government from the mid-1980s; Rahul Mukherji’s account of incremental idea-change	Cannot explain why reform waited for reserves to hit two weeks

• The strongest reading treats the crisis as the occasion rather than the cause. Shocks do not produce structural change; they destroy the political viability of the status quo long enough for a prepared alternative to be enacted. Reform requires both a crisis and a drawer with a plan in it.

The reform programme

Stabilisation and structural adjustment are two different things

The deepest change was in the state’s self-description. Before 1991 the state occupied the commanding heights and the private sector was deliberately kept subordinate; after 1991 the private sector was given those heights and the state redefined itself as facilitator and regulator. The single most common analytical error about 1991 is to treat the package as one thing. It was two, with different logics, horizons and distributive effects, and conflating them muddles every assessment of what worked.

	Stabilisation	Structural adjustment
Objective	Restore external and fiscal balance	Change what the economy produces and how
Instruments	Devaluation, fiscal compression, monetary tightening, expenditure cuts	Delicensing, trade opening, financial reform, competition policy
Horizon	Twelve to eighteen months	A generation
Distributive effect	Immediate and regressive — the poverty ratio rose in 1993-94	Diffuse gains, concentrated losses
Test of success	Reserves, deficit, inflation	Productivity, exports, entry of new firms

• The short-run poverty increase after 1991 was a stabilisation effect, not a verdict on liberalisation — the consequence of expenditure compression and a devaluation-driven price rise, and attributing it to structural reform confuses the medicine with the surgery.

The exchange rate, and the road deliberately not taken

• Two devaluations, on 1 and 3 July 1991, cut the rupee by about 18-19% cumulatively against major currencies. They were executed in two steps so that the government could stop after the first if the reaction proved unmanageable.
• The Liberalised Exchange Rate Management System (LERMS) of March 1992 created a dual exchange rate: 40% of export earnings surrendered at the official rate, 60% convertible at the market rate, unified into a single market-determined rate in March 1993.
• India accepted the obligations of Article VIII of the IMF’s Articles of Agreement on 20 August 1994, establishing current-account convertibility — free convertibility for trade, services, interest, dividends and remittances.
• Capital-account convertibility was deliberately refused. The Tarapore Committees of 1997 and 2006 each set a phased roadmap conditional on fiscal consolidation, low inflation and a cleaned-up banking system; the preconditions were never declared met.
• That refusal is arguably the single best decision of the reform era. Thailand, Indonesia and Korea, with open capital accounts and weak balance sheets, were devastated in 1997 by exactly the reversal India had made structurally impossible. The reform India did not do explains its resilience better than most of the ones it did.

Industrial policy: the New Industrial Policy of 24 July 1991

• Announced the same day as the budget, the New Industrial Policy abolished industrial licensing for all industries except a reserved list of eighteen, since pruned to a handful concerning security, hazardous chemicals, tobacco, explosives and defence aerospace.
• The MRTP asset threshold and the requirement of prior approval for expansion, merger and takeover were removed, and the Act was replaced by the Competition Act, 2002, with the Competition Commission of India operational from 2009. The shift is conceptually exact: from regulating size to regulating conduct — abuse of dominance, anti-competitive agreements, combinations above a threshold.
• Public-sector reservation shrank from 17 industries to a handful — atomic energy, atomic minerals and railway operations — with defence, telecommunications, air transport, power and petroleum all opened. As an instrument of reform, disinvestment moved from minority stake sales in 1991-92 to strategic sale, and was reframed by the Public Sector Enterprise Policy of 2021 around a bare minimum presence in four strategic sectors and exit elsewhere.
• Small-scale reservation was dismantled by attrition, not by decision. The list was cut in stages from the late 1990s and the last twenty items were de-reserved in April 2015.

Trade

• Peak tariffs fell from above 300% to 150% in 1991-92 and then in stages to a peak of 40% by 1997-98 and around 20% by the mid-2000s. The simple average applied tariff, over 100% in 1990, fell to roughly 12-13%.
• Import licensing for capital goods and intermediates went first, because they were inputs to production and the lobby against them was weak. Quantitative restrictions on consumer goods survived until 2001, removed only after India lost a WTO dispute brought by the United States — an instructive case of external obligation doing what domestic politics would not.
• The vocabulary changed with the substance: the EXIM Policy became the Foreign Trade Policy, signalling a shift from managing permitted transactions to promoting a trade strategy. The Special Economic Zones Act, 2005 created enclaves with tax and customs concessions and relaxed labour administration.
• Trade openness rose from about 15% of GDP in 1990 to a peak above 55% around 2012 before falling back, and services exports — software, business services, now global capability centres — became the distinctive success, in a sector the licensing regime had never learned to control because it did not exist when the rules were written.

Foreign investment

• The automatic route replaced case-by-case clearance: investment up to a sectoral cap needs only post-facto notification to the Reserve Bank, with the government route retained above caps and in sensitive sectors.
• FERA was replaced by the Foreign Exchange Management Act, 1999, and the change of a single word carries the argument: from regulation to management, and from a criminal offence to a civil contravention.
• The FDI/FPI distinction matters for stability. Foreign direct investment buys management participation and is illiquid; foreign portfolio investment buys listed securities and can leave in a morning.
• Current sectoral position: 100% automatic in most manufacturing, single-brand retail, contract manufacturing and coal mining; insurance raised to 100% in the 2025-26 budget from 74%; defence 74% automatic and 100% by government route; multi-brand retail capped at 51% with conditions and effectively dormant.
• The FDI record now diverges sharply depending on which number is used. Gross FDI inflow reached a record $94.84 billion in 2025-26, up from $80.61 billion the previous year. Net FDI, however, was only $6.95 billion — better than the $0.96 billion of 2024-25, but a fraction of the gross figure, because repatriation by foreign investors and outward investment by Indian firms have both surged. India is now simultaneously a leading destination and a leading exporter of capital, which headline gross-inflow figures conceal.

Financial-sector reform and the architecture built afterwards

• The Narasimham Committee of 1991 set the framework: cut the statutory liquidity and cash reserve ratios that pre-empted bank funds for government borrowing, introduce prudential norms for income recognition, asset classification and provisioning, adopt capital adequacy on Basel lines, deregulate interest rates, and admit private and foreign banks. The second Narasimham Committee of 1998 shifted to consolidation, stronger capital and the legal machinery for recovery.
• The Securities and Exchange Board of India became statutory in 1992 and the Controller of Capital Issues was abolished — firms could now price their own share issues instead of having the price fixed by an official.
• The later architecture is the second act of financial reform:
  • The Financial Sector Legislative Reforms Commission under Justice B.N. Srikrishna proposed an Indian Financial Code unifying a fragmented regulatory structure; its main recommendations were never enacted.
  • Inflation targeting was adopted through the RBI Act amendment of 2016, setting a 4% CPI target with a ±2% band, and creating a six-member Monetary Policy Committee with three external members.
  • The Insolvency and Bankruptcy Code, 2016 replaced a debtor-in-possession regime that took years with a creditor-in-control, time-bound resolution process.
  • The asset quality review of 2015 forced banks to recognise bad loans they had been evergreening, exposing the twin balance-sheet problem — over-leveraged corporates facing under-capitalised banks, neither able to fix itself without the other.

Fiscal and tax reform

• The Chelliah Committee of the early 1990s set the template still in use: lower rates, fewer rates, wider base, better administration.
• The Kelkar Task Forces of 2002 and 2004 carried this into administration and into the case for a comprehensive goods and services tax.
• State-level Value Added Tax replaced sales tax from 1 April 2005, ending the cascading of tax on tax within each state — the work of the empowered committee of state finance ministers, and the institutional ancestor of the GST Council.
• The Fiscal Responsibility and Budget Management Act, 2003 legislated deficit targets and mandated medium-term fiscal statements.
• GST came into force on 1 July 2017 under the 101st Constitutional Amendment, replacing seventeen central and state levies and creating the GST Council, India’s most consequential experiment in shared fiscal sovereignty. GST 2.0, effective 22 September 2025, collapsed the four-slab structure into two main rates of 5% and 18%, with a 40% demerit rate on a short list — closer to the single-rate ideal the original design abandoned for political reasons.
• Subsidy reform moved from price to person. Direct benefit transfer, built on the Jan Dhan–Aadhaar–Mobile trinity, replaced administered prices with cash in accounts for cooking gas, fertiliser, scholarships and pensions. It is the most successful administrative reform of the post-1991 period.

The politics of reform

This is where the Indian case earns its place in comparative political economy. The puzzle is not that India liberalised — most indebted developing countries did — but that a poor, unequal, federal, competitive democracy sustained an unpopular programme for three decades without any government ever winning a mandate for it.

Reform by stealth

• The phrase describes a deliberate technique: reforms were made in ways designed not to be noticed. Executive notification rather than legislation; administrative circular rather than order; announcements buried among many others on budget day; the industrial policy released the same afternoon as the budget so that it competed with the budget for attention.
• Nothing needed Parliament that could be done without it. Delicensing, tariff changes, the exchange rate and FDI caps could all be altered administratively. The reforms India did accomplish are almost exactly the set that required no legislative majority; land and labour, which require one, remain unfinished.
• Because reform was never argued for publicly, no constituency was built for it and the gains have no political owner.

Liberalisation succeeded politically by never asking the electorate’s permission, and that success is the source of every difficulty it has had since.

Gradualism against big bang

	India	Russia	China
Approach	Gradualist and sequenced	Shock therapy — simultaneous price liberalisation, privatisation, stabilisation	Dual-track — market prices at the margin, plan prices for quotas
Political setting	Competitive democracy, coalition governments	Post-collapse state	Single-party state
Ownership change	Slow, minority stakes first	Mass voucher privatisation	Township and village enterprises, not privatisation
Outcome in the 1990s	Growth accelerated modestly, no output collapse	Output fell about 40%, life expectancy fell, oligarchy formed	Sustained double-digit growth
Capital account	Kept closed	Opened	Kept closed

• Joseph Stiglitz’s critique of the Washington Consensus turns on sequencing: liberalising capital flows before building financial supervision, or privatising before building competition law and corporate governance, produces asset-stripping rather than efficiency.
• India’s gradualism was a political necessity that turned into an economic virtue. Coalition arithmetic, federal veto points and organised labour made incrementalism the only feasible speed, and that constraint prevented the sequencing errors Stiglitz identifies, kept the capital account shut, and made the reforms cumulative rather than reversible. Slowness also did political work: gradualism gave the consensus time to form rather than requiring it in advance.
• The cost is equally real. Gradualism protects incumbents. Sectors with organised losers — agriculture, labour law, land acquisition, public-sector employment, electricity distribution — were postponed indefinitely, and postponement in a democracy is rarely temporary.

The reforms India could make without Parliament were made; the reforms that needed Parliament are still waiting.

Democratic politics and economic reform: the Jenkins account

Rob Jenkins’s Democratic Politics and Economic Reform in India is the indispensable study of the question, because it takes seriously what most accounts treat as an obstacle — that India is a democracy — and shows how democratic institutions were used as instruments of reform rather than merely survived by it.

• Political skill, not insulation, is the explanation. Jenkins rejects the then-standard view that reform requires an autonomous technocracy shielded from politics. Indian reformers were not insulated; they were skilled at using the ordinary equipment of Indian politics — ambiguity, patronage, incrementalism and federal complexity — to advance an agenda they could not announce.
• Obfuscation was a governing technique. Reforms were disguised, renamed, presented as continuity with Nehruvian goals, and attributed to external compulsion when that helped and denied when it did not — the same measure described to different audiences in incompatible terms because no audience saw the whole.
• Federalism dispersed the blame and the burden. Politically costly measures — user charges, electricity tariffs, closure of loss-making units, labour flexibility in practice — devolved to states, so the Centre took the credit for growth while states absorbed the resentment. Federalism also allowed variation: a measure resisted nationally could be tried in one state.
• There was no losers’ coalition. This is Jenkins’s sharpest observation. The losses from reform were real but dispersed, delayed, and difficult to attribute — a worker who never got a job in a factory that was never built does not know to protest. The winners were concentrated and organised; the losers were diffuse, and where they were concentrated (organised labour, public-sector employees) they were protected precisely so that they would not organise. Job security in the public sector was the price of not opposing reform elsewhere.
• Informal institutions did real work. The lubrication of Indian politics — discretionary funds, transfer postings, party finance, the accommodation of rent-seeking — let politicians compensate those whose rents were disrupted, converting opponents into participants.
• The limitation of the argument is its own success: a reform sustained by concealment cannot be deepened by persuasion.

Varshney: elite politics and mass politics

• Ashutosh Varshney distinguishes the elite arena — parliament, bureaucracy, business associations, editorial pages — from the mass arena, where an issue enters street mobilisation, electoral campaigning and the vernacular press. His claim is that India’s reforms survived because they stayed in the elite arena.
• Trade policy, industrial licensing, capital markets and exchange rates are conceptually distant from everyday life; their effects are real but mediated. Voters do not experience a tariff; they experience a price.
• When reform crosses into the mass arena, it stops. Attempted privatisation of banks and insurance triggered general strikes; electricity tariff reform brought down state governments; land acquisition mobilised farmers; and the three farm laws of 2020 are the definitive test case — an attempt to extend liberalisation to agricultural marketing that entered the mass arena within weeks and was repealed in December 2021 before its provisions had been tested.
• The corollary is uncomfortable. Reform in India has been possible in proportion to its invisibility, so the reforms that matter most for the poor — agriculture, labour, land — are the ones the system finds hardest to make.

Kohli: a pro-business state, not a pro-market state

• Atul Kohli argues that India’s transformation is best described not as a shift from state to market but as a shift in the state’s alliance — from a state rhetorically committed to redistribution to a pro-business state allied with domestic capital.
• The distinction is analytically sharp and politically consequential:

	Pro-market	Pro-business
Beneficiary	Consumers and new entrants	Established incumbent firms
Attitude to competition	Welcomes it, including from imports	Welcomes deregulation, resists competition
Typical instrument	Antitrust, open entry, tariff reduction	Tax concessions, land at concessional rates, credit, protective tariffs
Attitude to labour	Flexible markets, portable protections	Suppression of organised labour
Test	Does a new firm find it easy to enter?	Does an existing firm find it easy to grow?

• Kohli locates the turn in the 1980s, under Indira and Rajiv Gandhi, which aligns him with Rodrik and Subramanian on timing while disagreeing about what the change signified. The state did not retreat; it changed sides.
• The evidence is not scarce: rising corporate concentration, protective tariffs persisting alongside deregulation, and the resilience of business houses across regime change.

Sinha, Mukherji, Nayar, Chatterjee and Bardhan

• Aseema Sinha shifts the unit of analysis downward, showing that the sub-national state was always a variable. Even under central licensing, states differed enormously in how they used their discretion — Gujarat’s bureaucracy brokered central approvals for its industrialists, West Bengal’s did not — and those pre-existing capacities, more than post-1991 policy, decided which states captured the gains.
• Rahul Mukherji argues for the primacy of ideas and incremental institutional learning. Technocrats absorbed the lessons of the failed 1966 devaluation and the partial successes of the 1980s, built a blueprint over two decades. On this reading 1991 was implementation, not conversion, and the externalist account mistakes the trigger for the author.
• Baldev Raj Nayar reads the reforms through economic nationalism rather than neoliberal capitulation. India liberalised to build national economic power, kept control of the capital account, resisted WTO pressure on agriculture and intellectual property, and never accepted the ideological package wholesale — which is why liberalisation coexists so comfortably with Atmanirbhar Bharat.
• Partha Chatterjee supplies the account of those outside the reform’s benefits. He distinguishes civil society — rights-bearing citizens who deal with the state through law — from political society, where the urban poor, encroachers and informal workers negotiate through claims recognised as political necessities rather than legal entitlements. Post-reform welfare is the governmentalisation of that negotiation:.
• Pranab Bardhan’s model of the dominant proprietary classes — industrial bourgeoisie, rich farmers, and the professional-bureaucratic elite — explained the pre-reform stalemate as the product of their competing claims on public resources, producing subsidies for all and investment for none. Updated for the post-reform period the coalition has shifted: industry gained decisively, the professional class gained through globalised services, and rich farmers were left holding subsidies rather than growth — the configuration that produced the farm agitation.

The consensus and its thinness

• Bimal Jalan’s observation stands: there is no longer a serious dispute over the desirability of reform, and every major party except the left now accepts that there is no alternative model on offer.
• The counter-observation is equally important. The content of the consensus is thin. It extends to macroeconomic stability, private enterprise, foreign investment and the abandonment of licensing; it does not extend to labour law, land acquisition, privatisation, agricultural markets, subsidy withdrawal or trade openness.
• Each party opposes in office what it proposed in opposition. The party that introduced disinvestment opposed its acceleration; the party that drafted the goods and services tax opposed it for years; the land acquisition law of 2013 passed with cross-party support and its dilution was attempted by the same benches two years later. The consensus is procedural rather than substantive — agreement that reform is necessary, without agreement on any specific reform.

Competitive federalism and the divergence of states

• Liberalisation devolved economic policy downward without formally amending the federal structure. Once the Centre stopped allocating licences, capital and foreign collaborations, everything that determined where investment went — land, electricity, water, labour administration, clearances, law and order — was a state subject. Delicensing was therefore an act of decentralisation disguised as deregulation.
• The “reform by competition” argument follows: states compete for mobile capital, and competition disciplines them into better regulation. The mechanism is real, but it is only as strong as a state’s initial capacity, and the weakest states cannot compete.
• The contrasting models are instructive rather than rankable. Gujarat built growth on infrastructure, land availability, port-led logistics and an administration that clears projects fast, with strong output growth and comparatively weak social indicators. Tamil Nadu reached comparable prosperity through early and sustained investment in schooling, health, nutrition and urbanisation, a broad industrial base and high female workforce participation. Kerala shows a third route in which social outcomes preceded income entirely.
• The evidence on convergence is negative. Studies of state per capita income since 1991 find sigma-divergence — dispersion across states has widened — and unconditional beta-divergence, meaning poorer states have not grown faster.
• The political cost of exclusion is measurable. The regions where growth arrived last and least are the districts where left-wing extremism took root — which is why the diagnosis that India’s economic crisis is at bottom a governance crisis, requiring economic and political reform together, has outlasted every particular scheme.
• The reasons are structural. Investment goes where infrastructure, skills, ports, markets and administrative reliability already are, so the returns to liberalisation are increasing in pre-existing capacity and opening the economy amplified the differences it inherited.
• The divergence has now become a constitutional question. Southern and western states have lower fertility, higher per capita income and a shrinking share of the national population; northern states the reverse. Because Lok Sabha seats have been frozen on the 1971 census, delimitation on current population would transfer parliamentary weight from the states that performed best on development to those that performed worst, and fiscal devolution formulas raise the same dilemma. Uneven development has generated a federal crisis that no economic instrument can resolve.

Crony capitalism and the reform backlash

• The critique is not that liberalisation created corruption but that it changed the location of the rent. The licence raj created rents by restricting entry, and abolishing licences removed them. But the state kept monopoly control over natural resources — spectrum, coal, minerals, land — and had no rule for allocating them, so the rent migrated from the permission to the resource.
• The 2G spectrum allocation used a first-come-first-served procedure at 2001 prices in 2008; the Comptroller and Auditor General’s presumptive-loss estimate became the defining number of the period, and the Supreme Court cancelled 122 licences in 2012. Coal block allocation by screening committee produced a comparable finding, and the Court cancelled 204 allocations in 2014.
• The remedy was procedural and it worked. Auctions were mandated for spectrum and coal, and the discretionary allocation of these resources effectively ended.
• The lineage of the diagnosis is older. The Vohra Committee report of 1993 on the nexus between crime, politics and bureaucracy described exactly the fusion of political finance and economic favour that the resource scandals later illustrated; its recommendations were never implemented.
• Political finance remains the unreformed core. The Electoral Bond Scheme of 2018 permitted anonymous, unlimited corporate donations and removed the cap restricting companies to 7.5% of average net profits. In Association for Democratic Reforms v. Union of India (February 2024) a five-judge Constitution Bench struck it down as violating the voter’s right to information under Article 19(1)(a), holding that anonymity was disproportionate. The judgment removed the instrument without replacing the demand it served.
• The political consequence matters as much as the legal one. The corruption backlash of 2011-14 was the closest liberalisation came to entering the mass arena, and its target was never liberalisation as such but its distributive unfairness, which is why it produced better allocation rules rather than reversal.

Democracy and development: the argument underneath

The reform debate sits on top of an older question about whether India’s political form helps or hinders its economic transformation. Both sides of that argument have serious versions, and neither has won.

• The “cruel dilemma” thesis, associated with Jagdish Bhagwati’s early work, holds that a democracy must satisfy consumption demands immediately while development requires deferring consumption to invest. Authoritarian regimes can impose the sacrifice; democracies must bid for consent, so they under-invest, over-subsidise and reform slowly.
• The “developmental state” counterfactual sharpens this. South Korea, Taiwan and Singapore industrialised under authoritarian governments with insulated technocracies that could discipline capital — withdrawing support from firms that failed to export — and suppress labour. India could do neither. Democracy prevented India from being Korea.
• Amartya Sen’s counter is the most powerful argument on the other side. No substantial famine has ever occurred in a functioning democracy with a free press and opposition parties. India’s last great famine was in Bengal in 1943, under colonial rule; China’s Great Leap famine of 1958-61 killed tens of millions in a system with no mechanism for the information to travel upward. Democracy is a protection against catastrophic policy error, and the value of that protection never appears in a growth rate.
• Sen’s second point is about what democracy did not do. The same democratic mechanisms that prevent famine have proved feeble against endemic undernutrition, illiteracy and preventable morbidity, because these produce no sudden, reportable event for a free press to respond to. Democracy handles crises well and chronic deprivation badly — which is India’s record.
• The cross-national evidence is genuinely indeterminate. Adam Przeworski’s large-sample work finds no systematic growth advantage for either regime type: authoritarian regimes have both the highest and the lowest growth rates while democracies cluster in the middle, so regime type predicts variance, not the mean.
• The Indian resolution is that democracy determined the method rather than the outcome. Reform had to be gradual, disguised, federally dispersed and compensated with welfare.

From ascriptive to developmental politics?

• The claim is that post-liberalisation Indian voters increasingly choose on the basis of delivery — roads, electricity, water, houses, gas connections, cash transfers — rather than caste and community. Several strands of evidence support it.
  • Bihar between 2005 and 2010 is the standard case: a government elected on a governance platform, delivering roads, school attendance, law and order and bicycles for girls, was returned with a larger majority.
  • The “new welfarism” literature of Abhishek Anand, Vikas Dhingra and Arvind Subramanian identifies the mechanism: the post-2014 model provides subsidised private goods at scale — toilets, cooking gas, bank accounts, electricity, housing, piped water — rather than public goods such as health and education.
  • Labharthi (beneficiary) politics is the political form this takes: a direct, named relationship between the scheme and the recipient, bypassing intermediaries. Direct benefit transfer is the technology that makes ascriptive intermediation unnecessary.
• The counter-evidence is strong and should not be minimised: ascriptive mobilisation has not receded, it has been reorganised.
  • Caste remains the primary axis of candidate selection, coalition arithmetic and vote transfer in most states, and the demand for caste enumeration and expanded reservation has intensified rather than faded.
  • Religious mobilisation has expanded, not contracted, over exactly the period of highest growth.
  • The two are complementary rather than sequential. The most successful contemporary formula combines broad welfare delivery with consolidated identity mobilisation — benefits reach beneficiaries across castes while the political appeal is made in identity terms.
  • Rajni Kothari’s older insight anticipates this: politics does not replace caste, it politicises caste, converting it from a ritual order into an instrument of secular claim-making.
• The defensible conclusion is that developmental politics has been added to ascriptive politics, not substituted for it. A party that delivers nothing can no longer win on identity alone; a party that delivers everything and offers no identity appeal has not yet shown that it can win either.

The record, argued

Growth, and why the number is contested

• The acceleration is real and large. From about 3.5% for three decades to the mid-1970s, growth rose to roughly 5.6% in the 1980s, 6% in the 1990s, and a boom of about 8-9% between 2003-04 and 2007-08 — the fastest sustained expansion in Indian history.
• The post-2011 record is weaker than the headline suggests. Growth decelerated through the mid-2010s; demonetisation in November 2016 and the transition to GST in 2017 were both disruptions concentrated on the informal and small-enterprise sector; COVID-19 produced a contraction of about 5.8% in 2020-21, among the deepest of any large economy, followed by a strong rebound.
• The current position is strong. India is the fourth-largest economy by nominal GDP and the third-largest by purchasing power parity, with a middle class large enough to be a market in its own right; growth has been in the 7-7.6% range across the last three years, and the new GDP series with base year 2022-23, released in February 2026, puts 2023-24 at 7.2% and 2024-25 at 7.1%, with 2025-26 estimated higher still.
• The measurement caveat is not pedantry. The 2015 revision to the 2011-12 base, using the MCA-21 corporate database, raised measured growth substantially. Arvind Subramanian, the government’s own former Chief Economic Adviser, argued that growth between 2011-12 and 2016-17 had been overstated by about 2.5 percentage points a year; the Economic Advisory Council to the Prime Minister rejected the methodology, and the back-series for pre-2011 years produced two incompatible sets of official numbers. The direction of the acceleration is not in doubt; its magnitude in specific sub-periods is.

How open is India actually? A four-way audit

• An Economic Survey audit of the transition measured India against four standard benchmarks rather than against rhetoric.
  • Openness to trade: relative to the size of its economy, India trades more with the world than a country of its size and income would be predicted to.
  • Openness to foreign capital: despite retained capital controls, net foreign capital inflow is normal for an emerging economy, with FDI at rates not far short of what China received at the height of its boom.
  • Extent of the public sector: against the myth of an unusually large state footprint, India has rolled the public sector back substantially — in civil aviation, telecommunications and financial services above all — while China has retained and expanded its state enterprises.
  • Government expenditure: total government spending is no larger than is normal for a country at India’s level of development.
• The Survey’s own verdict is double-edged: the transition is remarkable because it was achieved under an extremely competitive democratic system, — in civil aviation, telecommunications and financial services above all — while China has retained and expanded its state enterprises. India’s problem is not too little market or too much state; it is a state that opened well and delivers badly.

Poverty

Measure	Line	Latest estimate	What it captures
Tendulkar	Official, 2011-12	21.9% (2011-12)	The last official consumption-poverty headcount; no official line has been set since
Rangarajan	Higher line, 2011-12	29.5% (2011-12)	Never formally adopted
World Bank extreme poverty	$3.00/day, 2021 PPP	5.3% (2022-23)	Destitution; India now below the global average of about 10.8%
World Bank lower-middle-income line	$4.20/day, 2021 PPP	about 24% (2022)	A more demanding threshold — roughly a quarter of the population
NITI Aayog multidimensional	Health, education, living standards	11.28% (2022-23)	Deprivation beyond consumption

• Montek Singh Ahluwalia’s retrospective assessment is the standard practitioner’s account, and it is candid about the sequence. Poverty was 44.5% in 1983; it rose marginally in 1993-94, immediately after stabilisation; it fell from 45.3% in 1993-94 to 37.2% in 2004-05 on the Tendulkar line, but with absolute numbers still rising because population grew faster than the ratio fell; and then fell sharply to 21.9% in 2011-12, the first period in which the absolute number of poor also declined, from about 407 million to 270 million.
• Ahluwalia’s qualification is the important part: 270 million is not a success, and the population just above any poverty line is scarcely different from the population just below it. India’s record on basic services — schooling, health, sanitation and drinking water — remained poor throughout the period of fastest poverty decline — which is the strongest evidence that consumption poverty and deprivation are not the same variable.
• The choice of line drives the headline, and saying so is the analytical point rather than an evasion. On the $3.00 line India has nearly eliminated extreme poverty; on the $4.20 line a quarter of India is poor; on the multidimensional measure about one in nine.

Inequality: a live dispute, not a settled indictment

• The World Inequality Lab’s work by Nitin Kumar Bharti, Lucas Chancel, Thomas Piketty and Anmol Somanchi is the strongest case for the prosecution. Building an income and wealth series for 1922 to 2023, they find the top 1% income share at 22.6% and wealth share at 40.1% in 2022-23. They call the result the “Billionaire Raj” and propose a 2% super-tax on the net wealth of the 167 wealthiest families, yielding an estimated 0.5% of national income for health, education and nutrition.
• The consumption evidence points the other way. The Household Consumption Expenditure Survey of 2023-24 records the consumption Gini falling to 0.237 in rural India (from 0.266) and 0.284 in urban India (from 0.314), with the urban-rural gap in monthly per capita expenditure narrowing from 84% in 2011-12 to 70% in 2023-24.
• The divergence is real and has an explanation. Income and wealth are concentrating at the top while consumption levels at the bottom: consumption is bounded, transfers in kind raise measured consumption without touching income or wealth, and surveys miss the very rich. Both findings can be correct: the floor has risen and the ceiling has risen faster.
• A serious official position argues that this is the wrong thing to worry about at India’s stage of development. An Economic Survey chapter on inequality and growth found that in advanced economies inequality correlates with worse health, education, life expectancy, infant mortality and crime outcomes while per capita income does not — but that across Indian states both inequality and per capita income correlate similarly with those outcomes, so the conflict observed in rich countries does not appear in India. It also found a strong negative relationship between state income and poverty and none between inequality and poverty, concluding that growth does far more for the poor than redistribution at India’s income level. It draws on Martin Feldstein’s argument for targeting poverty rather than inequality, and against John Rawls’s maximin criterion cites experimental evidence that people in the original position choose to maximise average income subject to a floor.
• That argument deserves to be taken seriously and its limits stated. It rests on cross-sectional correlations rather than causal identification; states that invested in capabilities first may simply have grown faster afterwards, which reverses the arrow; and it answers the instrumental case against inequality while leaving the political case untouched. Concentrated wealth buys political influence, and a correlation between state income and health outcomes says nothing about that.

“Poverty is the parent of revolution and crime.” — Aristotle

Employment: the sharpest failure

• The core charge is that growth has been jobless — output has risen much faster than employment, and the transition that every successful industrialiser has made, from agriculture into factory work, did not happen.
• Manufacturing never took the share it was supposed to. Its share of value added has been stuck around 16-17% for three decades against the 25% target of Make in India (2014), and the World Bank’s measure of manufacturing in GDP fell from 15% in 2018 to 13% in 2024. India moved from agriculture directly into low-productivity services and construction — premature deindustrialisation.
• The headline unemployment figures look benign and mislead. The Periodic Labour Force Survey for 2025 records an unemployment rate of 3.1% on usual status for those aged 15 and above, labour force participation of 59.3% and a worker-population ratio of 57.4%. In an economy with almost no unemployment insurance, a 3% unemployment rate means people cannot afford to be unemployed, not that work is plentiful.
• The composition is the problem. Roughly 56% of workers are self-employed, with a large share of those unpaid helpers in family enterprises; regular wage employment is about 24%; and a majority of even regular workers have no written contract and no social security.
• Female participation rose sharply, and the reason is contested. Female labour force participation for those aged 15 and above fell to a low near 23% in 2017-18 and has since risen to 40%, with rural female participation at 45.9%. Work by Maitreesh Ghatak, Mrinalini Jha and Jitendra Singh finds that real average daily earnings of own-account workers fell about 8% between 2017-18 and 2023-24, and that unpaid family helpers effectively earn around ₹50 a day — under a fifth of what the lowest-paid casual workers earn. Much of the rise therefore looks like distress participation: households pushing extra members into low-return self-employment because primary incomes fell.
• The jobs that were created were largely bad ones. Wages stayed low, the work available was overwhelmingly unskilled by comparison with other Asian economies at the same stage, and the export enclaves built to attract investment reproduced the problem they were meant to solve: conditions in special economic zones have borne hardest on women workers.
• Workers returned to agriculture after 2019, reversing the structural direction of the previous decade. A developing economy in which the farm sector re-absorbs labour is not industrialising.
• Statistical discontinuity complicates every comparison. The PLFS replaced the quinquennial NSS Employment-Unemployment Survey in 2017-18, changed the sampling design and periodicity, and moved to monthly frequency from January 2025. Series that are not strictly comparable are routinely compared, and the direction of any conclusion depends on which vintage is used.
• The gig economy is the fastest-growing form of new work and the least protected. NITI Aayog projects about 2.35 crore gig workers by 2029-30. The Code on Social Security, 2020 was the first Indian statute to define gig and platform workers and to provide a social security fund financed by aggregator contributions of 1-2% of turnover, with portable Aadhaar-linked registration through e-Shram.
• The failure has a clear cause. Labour-intensive manufacturing was the missing channel, blocked by small-scale reservation until 2015, rigid dismissal law above the 100-worker threshold, unreliable power, poor logistics and an exchange rate that never favoured exports as East Asia’s did. This is the specific sense in which liberalisation was not accompanied by adequate reform: the product market was freed and the factor markets were not.

Human development and the Sen–Bhagwati debate

• Amartya Sen’s central charge is that the growth pattern matters more than the growth rate, and that liberalisation produced growth without a corresponding transformation in living standards. Progress on several social indicators slowed relative to the pre-reform decades; agriculture stagnated; and the services boom did not reach the bottom because people lacked the education and health to enter it. Deprivation deepened in the BIMARU states — Bihar, Madhya Pradesh, Rajasthan and Uttar Pradesh — and hunger persisted amid full grain warehouses, a failure of public action rather than of production.
• The China comparison is his sharpest instrument. China entered its market reforms of 1978 with near-universal basic schooling and a functioning rural health system, and could convert opening into mass employment. India attempted the same transition with an under-educated and under-treated workforce, and got services-led growth that most of its people could not enter. Bangladesh, poorer than India, has overtaken it on life expectancy, child immunisation and female schooling.
• The Uncertain Glory argument, made with Jean Drèze, is that India has grown rich in parts while remaining a laggard in the basic capabilities that make growth worth having, and that this is a failure of public services rather than of scarcity.
• The spending record supports the charge. Government health expenditure remains around 1.8-1.9% of GDP against the National Health Policy target of 2.5%, and education around 2.7-4% of GDP against the 6% target official since 1968 and restated in the National Education Policy 2020 — both below the lower-middle-income average.
• The counter-case from Jagdish Bhagwati and Arvind Panagariya is not a denial of the deprivation. Their argument is about sequence and feasibility: growth is the precondition of the fiscal resources that fund social spending, the states that grew fastest expanded social provision fastest, and pre-1991 India showed conclusively that redistribution without growth redistributes poverty.
• Sen’s rejoinder is that the causation runs both ways and the sequence is not optional. Education and health are inputs to growth, not merely its rewards; Waiting for growth to fund capability while capability constrains growth is a trap, not a sequence.
• The debate is the defining Indian argument about what reform was for.
• Bimal Jalan’s assessment, in Emerging India: Economics, Politics and Reforms, is the most balanced statement of the paradox: the reforms released creative energy and accelerated growth and productivity, and India’s record on social indicators remains among the worst in the world. His conclusion is not that reform failed but that it would be a grave mistake to leave the critical reforms outstanding.

	Sen and Drèze	Bhagwati and Panagariya
Binding constraint	Human capability — health, schooling, nutrition	Growth and productive employment
Reading of 1991	Necessary but insufficient; the social half was never done	Necessary and vindicated; incomplete because reform stopped
Model to emulate	China’s pre-reform capability base; Kerala; Bangladesh on social indicators	Gujarat’s growth model; East Asian export manufacturing
Policy priority	Public services, universal provision, public action	Deeper reform in labour, land, trade; growth-financed transfers
Where they agree	Both regard the current level of Indian social spending as inadequate

Agriculture: the sector reform did not reach

• Agriculture was left outside the 1991 programme almost entirely. Licensing, tariffs, foreign investment and the financial sector were reformed; the Agricultural Produce Market Committee system, the Essential Commodities Act, the minimum support price regime, input subsidies and land leasing restrictions were not.
• Such liberalisation as reached agriculture came through trade, not domestic market reform — quantitative restrictions removed under WTO obligation by 2001, tariff bindings, and exposure to world price volatility. Farmers received the risks of openness without the benefits of a freer domestic market, which is why agrarian distress and liberalisation are tightly linked in the rural political imagination.
• The terms of trade moved against agriculture through much of the reform period, and the sector’s share of GDP fell to around 16-18% while it continued to support over 45% of the workforce.
• The three farm laws of 2020 were the one serious attempt to extend reform to agriculture, and their fate is the clearest demonstration of Varshney’s argument in the record. Enacted by ordinance without consultation, they produced a year-long farmer mobilisation and were repealed in December 2021. The legal guarantee of a minimum support price remains the central unmet demand.
• A reform that redistributes risk cannot be introduced by stealth, because the losers can identify themselves.

What liberalisation did to the state

• Globalisation does not abolish the state; it reallocates what the state can decide.
• Anthony Giddens’s definition frames the mechanism:

“Globalisation can be defined as the intensification of worldwide social relations which link distant localities in such a way that local happenings are shaped by events occurring many miles away and vice versa.” — Anthony Giddens

• The regulatory regimes of the IMF, the World Bank and the WTO function as a de facto layer of global governance, setting obligations national law must accommodate. India’s patent regime under TRIPS, the removal of quantitative restrictions and the subsidy disciplines on agriculture are all instances of domestic legislative choice narrowed by external obligation.
• Multinational corporations shape policy options without eliminating sovereignty. They still need the state’s permission to enter and its stability to invest; what changes is that the state now competes for their presence rather than rationing it.
• The process is dualistic: it integrates and fragments at once, strengthening the states that write the rules and weakening those that receive them.
• The most useful concept for Indian politics is the democratic deficit this creates. In developing countries, expectations of the state are high and state capacities are low, and liberalisation widens exactly that gap: citizens continue to hold national governments accountable for outcomes those governments no longer control. Economic nationalism is the predictable political product of economic globalisation, not its opposite.

India freed the market for goods and left the markets for land and labour close to where it found them.

The unfinished agenda

The 2020 charge — that liberalisation was not accompanied by adequate reforms — is best answered item by item, because the gaps are specific rather than general.

Factor markets: land and labour

• Land acquisition is governed by the Right to Fair Compensation and Transparency in Land Acquisition, Rehabilitation and Resettlement Act, 2013, which replaced the colonial 1894 Act and introduced consent requirements (70% for public-private partnership projects, 80% for private), social impact assessment, and compensation at two to four times market value. The 2015 ordinance to dilute consent and assessment was promulgated three times and abandoned, and the subject reverted to the states. India now has neither a fast acquisition process nor a functioning land market, because titles remain presumptive rather than conclusive.
• Labour law consolidation is the flagship of second-generation reform, and its history illustrates the pattern exactly. The four labour codes — Code on Wages 2019, Industrial Relations Code 2020, Code on Social Security 2020, and Occupational Safety, Health and Working Conditions Code 2020 — consolidate 29 central labour laws. They were passed between 2019 and 2020 and brought into force only on 21 November 2025, a delay of five years.
  • The substantive changes are real: the retrenchment and closure threshold raised from 100 to 300 workers, a statutory national floor wage, a 50% cap on allowances outside the wage definition for computing provident fund and gratuity, pro-rata gratuity for fixed-term employees, and the first statutory recognition of gig and platform workers.
  • Operation remains partial. Central rules were still in draft into 2026; only about eleven states had notified final rules under at least one code, and the floor wage, gig-worker scheme and inspection framework remained unnotified. Because labour is a Concurrent List subject, a code in force centrally is not a code in operation nationally.
  • Central trade unions and farmer organisations have jointly demanded repeal, and the contest over the codes is live rather than settled.

The rest of the list

Area	What is unfinished	Why it has not moved
Agricultural marketing	Market committee monopoly, Essential Commodities Act, restrictions on leasing	Mass-arena politics; the 2021 repeal set the precedent
Fiscal structure	Subsidy composition, low tax-to-GDP ratio, states’ committed expenditure	Every subsidy has an organised recipient
Privatisation	Repeated slippage against targets; IDBI Bank’s strategic sale stalled and reported paused in 2026	Employee resistance, valuation disputes, political cost
State capacity	Regulatory quality, staffing of tribunals, technical capability of regulators	No electoral constituency for administrative reform
Contract enforcement	Judicial delay; commercial disputes measured in years	Requires judicial reform, which is outside executive control
Human capital	Health and education spending far below targets; learning outcomes	Long payback periods; benefits accrue to a later government

• Montek Singh Ahluwalia’s own list is the most authoritative statement of the agenda from inside it: tackle corruption; end red tape; invest in human capital; reform the factor markets in land and labour; expand job opportunities; and rehabilitate state capacity. Only one of the six is a market-opening measure — the unfinished agenda is overwhelmingly about the quality of the state, not the extent of the market.
• Ahluwalia’s arithmetic gives the stakes: reducing poverty to genuinely marginal levels requires sustained growth of about 8% for two decades, unattainable with an unskilled and unhealthy workforce.
• The “second-generation reforms” vocabulary has stayed a slogan for a reason. First-generation reforms were subtractive — they removed a licence, a quota, a control — and could be executed by notification. Second-generation reforms are constructive: they require building regulators, courts, schools, land records and inspection systems that work. You cannot deregulate your way to state capacity.
• The middle-income trap is the standing risk. India has escaped low-income status and now faces the transition most middle-income countries fail — wages too high to compete on cost, productivity too low to compete on quality. The demographic dividend window closes around the 2040s, and a dividend requires employed workers; an unemployed young population is a liability. India is currently spending the window rather than using it.

Where the argument stands now

• Industrial policy has returned, and nobody calls it that. The Production Linked Incentive scheme, launched in 2020 across fourteen sectors with an outlay of about ₹1.97 lakh crore, pays firms for incremental output — a targeted subsidy to industries chosen by the state, which is what industrial policy is. Its record, as of late 2025: realised investment of about ₹2 lakh crore against a ₹3.48 lakh crore target, incremental production of ₹18.7 lakh crore against ₹29 lakh crore, 12.6 lakh jobs against a projected 39 lakh, and incentives actually disbursed amounting to only about 12% of the outlay. Electronics is the clear success, with production up 146% between 2020-21 and 2024-25, while battery cells and solar have badly underdelivered. The scheme has produced assembly, not yet an ecosystem.
• Tariffs have been rising since 2018, reversing three decades of reduction, and the charge of “premature protectionism” is that India is closing before it has industrialised, raising input costs for the very exporters it wants to build. The defence is that the simple average applied tariff of about 13.4% overstates the reality, since the trade-weighted average is near 7%.
• Atmanirbhar Bharat is not straightforwardly a return to import substitution. The classical version restricted imports to protect a domestic market; the current version subsidises production for export competitiveness. But the instruments — tariffs, local-content requirements, subsidies to chosen firms — are the same, and protection granted for a transition tends to outlive the transition, because the protected acquire an interest in it.
• Digital public infrastructure is the genuinely novel Indian contribution, and the strongest claim that India has produced a development model rather than adopted one. Aadhaar, the Unified Payments Interface, Jan Dhan accounts, the account aggregator framework and the Open Network for Digital Commerce are open, interoperable public rails on which private services compete — lowering transaction costs at population scale without either state ownership or private monopoly.
• The trade posture has changed decisively. After walking out of the Regional Comprehensive Economic Partnership in 2019, India pursued bilateral agreements instead: the UAE CEPA (2022), Australia ECTA (2022), the EFTA Trade and Economic Partnership Agreement (2024) with its $100 billion investment commitment, the India-UK Comprehensive Economic and Trade Agreement signed 24 July 2025, agreements with Oman (December 2025) and New Zealand (December 2025), and an announced India-EU agreement (January 2026). The strategy is bilateral and selective — openness to partners chosen individually rather than to a bloc that includes China.
• Welfare through direct transfers is the political settlement that makes reform survivable. Free foodgrain to over 80 crore people, cash transfers to farmers and women, subsidised housing, cooking gas and health insurance are not a departure from the reform model; they are its precondition, supplying the compensation that lets a liberalised economy retain electoral consent.
• Is India reforming, consolidating or drifting? All three readings are defensible and the honest answer distinguishes domains. GST 2.0, the insolvency code, inflation targeting and the labour codes are real reform. The digital rails, the transfer architecture and the deregulation agenda are consolidation. Land, agricultural marketing, privatisation, state capacity and human capital are drift. The Economic Survey’s own framing — a shift from “Ruler’s Raj to Citizen’s Raj”, deregulation as institutional reorientation rather than state withdrawal, and state capacity as the binding constraint — is an accurate diagnosis; whether it is a programme or a description is the open question.

Conclusion

The reforms of 1991 did what they were designed to do and were never designed to do what India most needed. They removed a system of permissions that had constrained enterprise for four decades, ended the foreign-exchange constraint that had determined policy since independence, and produced three decades of growth that lifted more people out of destitution than any comparable period in Indian history.

• What they did not build is the schools, hospitals, land records, courts and labour institutions on which the next transition depends, because those require the state to be constructed rather than removed.
• The technique that made liberalisation possible — quietly, without a mandate or a public argument — is useless for building anything, which is the deepest reason the second-generation agenda has stalled.
• The debate is therefore no longer about whether the cage should have been opened. It is about the condition of the animal that came out.