Manish Sabharwal, Shruti Rajagopalan - "Deregulating the Economy" | 6th GW India Conference

ELI5/TLDR

Two of India’s sharpest policy minds explain why India still has 45% of its workforce on farms 35 years after the 1991 reforms. Manish Sabharwal argues the binding constraint is “regulatory cholesterol” — 67,000 compliances, 26,000 with jail terms, plus a bureaucracy that reads “prohibited till permitted.” Shruti Rajagopalan walks through two unsexy laws — agricultural land ceilings and land conversion rules — that together trap Indian farmers on plots smaller than American backyards and make turning a farm into anything else a multi-year bureaucratic obstacle course. Both see real momentum in 2026 (Jan Vishwas 2.0, labor codes, decriminalization), but warn that without civil service reform the flow of new compliance will refill whatever stock is cleared.

The Full Story

The frame: deregulation is the cake, not the sprinkles

Rick Rosso, the moderator, opens with a useful reframe. The headlines — tariffs, trade deals, immigration — are the sprinkles. The cake is regulation. Whether a business can actually produce something, grow, and take off is decided in Delhi, Bhubaneswar, Bhopal, not Washington or Brussels. India just grew trade by 13% in a tariff year, so external pressure has eased. The question is whether the domestic reform momentum survives that comfort.

Sabharwal: regulatory cholesterol

Manish Sabharwal — co-founder of TeamLease, India’s largest staffing firm, ex-RBI board — admits he is disillusioned with monetary and fiscal policy as engines of mass prosperity. Government spending has gone up 108x since 1990; per capita income only 8x. Public debt is at 80-90% of GDP. If fiscal deficits made countries rich, nobody would be poor. So what is the binding constraint?

His answer: regulatory cholesterol. After four years of database work, his team handed the government a list of 67,000 compliances, 6,700 filings, and 26,000 ways to go to jail. He breaks the pathology into six blind spots:

The compliance blind spot. Government reads acts and rules and stops. But the executive has invented 17 other instruments — guidelines, circulars, notifications, office orders, government memoranda — each of which can mandate behavior and punish non-compliance. The Constitution authorized one level of subsidiary legislation; the administrative state has built seventeen.
Excessive criminalization. The Ministry of Corporate Affairs has a 10-year jail provision for fraud — and 128 separate two-year jail provisions for stupidity (filing a form late, getting a meeting date wrong). India is “probably the only country in the world” to make ticketless rail travel a jailable offence; out of 800 crore travelers last year, eight cases were filed. Check bouncing as a criminal matter accounts for 15% of India’s entire court pendency — 4.3 million cases.
Prohibited till permitted. “Who allowed you to do this?” — a question Sabharwal says he has been asked 108 times by bureaucrats. He calls it profound from a theory-of-knowledge standpoint.
Instrument proliferation. See above.
Enforcing the unenforceable. Corruption, in Sabharwal’s definition, is the transmission loss between how a law is written, interpreted, practiced, and enforced. The wider that gap, the more “show me the person, I’ll show you the rule” becomes the operating system.
No single source of truth. India still has no canonical, machine-readable index of its own laws. The Supreme Court has said silent or unnotified instruments cannot be applied — but without a registry, no one can prove what is silent.

His three-D fix: decriminalize, deregulate, digitize.

On decriminalization, the first attempt (Jan Vishwas 1.0) asked each ministry to surrender its own jail provisions — “nobody cuts the tree they’re sitting on.” That removed 50 of 26,000. The new approach writes principles (does the provision protect against harm to others; is it already covered elsewhere) and applies them ministry by ministry, vetted by a special committee. Jan Vishwas 2.0 plus the labor codes plus 100 repealed laws plus the Companies Act decriminalization removes about 12,500 compliances. Tax decriminalization didn’t make it through — revenue authorities pushed back.

On deregulation, the new principles say: licenses outside four areas (national security, public safety, human health, environment) should be replaced by self-registration; licenses should be perpetual; inspections should be random and risk-based.

On digitization, four cabinet-approved projects: a single employer ID (replacing 26 different numbers), an enterprise version of DigiLocker, an API framework instead of PDFs, and “India Code” — a website where, by mid-2027, if a law is not on it, you do not have to comply with it.

His honest caveat: deregulation without civil service reform is a fool’s errand. Even if you clear the stock of compliance, the 25 million civil servants who execute the system will keep generating flow. But that reform’s time has not come; this one’s has.

Rajagopalan: the unsexy laws that trap Indian agriculture

Shruti Rajagopalan — Mercatus Center, NYU Classical Liberal Institute, host of Ideas of India — promises she will bore the audience to death by talking about land ceiling laws and land conversion laws. She mostly succeeds at the first part and entirely succeeds at the more important point.

Structural transformation is the textbook script: agricultural productivity rises, farms become factories, villages become cities. Land is the essential input and the essential bottleneck — immobile, inelastic, easily trapped by bad rules.

Land ceiling laws put a literal roof on how much agricultural land one person or family can own. Cross it and the government takes the surplus. Most Indian states have at least one ceiling law; many have several. Ceilings are tighter on irrigated land than dry land, because the original goal was redistribution, so the perverse intended consequence is that India has made its most naturally productive land the least productive. Combined with the absence of primogeniture, plots shrink with every generation. Average operational holding size in India: 1.08 hectares. In Kerala: 0.18 hectares — smaller, Rajagopalan notes, than her neighbor’s backyard in Falls Church.

The result on the ground: India’s cereal yields sit below Bangladesh, below Sri Lanka, just above the African continental average. Selling to a neighboring farmer often pushes that farmer over the ceiling, triggering confiscation. Selling to a non-farmer is illegal in many states. Government investment in irrigation can push you over the irrigated-land ceiling and trigger confiscation. To paper over the absurdity, states have layered exemptions for sugar lobbies, plantation lobbies, educational institutes, stud farms, and APMC markets — “a nice exercise in public choice.” Then, to prevent fragmentation getting worse, states added consolidation laws with minimum-floor sizes. Ceilings on top, floors below, only the number of children unknown.

Land conversion laws govern the move from agricultural to non-agricultural use. Rajagopalan’s team studied 50+ statutes across 20 states. The pipeline:

Documentation — spelling mistakes, missing provenance, dependent on state capacity.
Pre-scrutiny / tenure check — are you over the ceiling, in a forest area, in a PESA tribal area.
Site inspection — “the fertile ground for rent-seeking,” when officials show up. Worse: in UP and Bihar they often don’t show up at all, for months.
Inter-departmental clearances — forest, railways, irrigation, etc., in no published order. “Where things go to die.”
Fee assessment and mutation — Maharashtra used to charge up to 75% of land value to convert.
Government review — usually no time-bound deemed approval.
Compliance — some states will revoke your conversion certificate if you don’t actually start the new use within a year, even if the next-stage permit is the bottleneck.

Kerala is the case study from hell. To convert wetland from paddy to rubber requires a near-unanimous committee vote. Refuse to grow paddy and the state can requisition the land and hand it to someone who will. There are criminal provisions. “No wonder everyone’s fleeing to the Gulf.”

Her killer point as a law-and-economics scholar: she could find no externalities these conversion laws are addressing. Any real environmental or fire-safety concern gets handled at the next stage when an actual factory or apartment block is permitted. The conversion regime is a freestanding rent-extraction layer with no benefit side of the cost-benefit ledger. Her recommendation: just repeal it. Andhra Pradesh has, and Maharashtra has gone partway.

The diagnosis problem

Rajagopalan flags an intellectual difficulty with Sabharwal’s framework: the three-D approach assumes you know what to deregulate. With land conversion, every state has different binding constraints — paperwork in one, inspection capacity in another, fees in a third — and removing one just exposes the next. So the answer is not surgical decriminalization but wholesale repeal where there’s no benefit to weigh.

Labor codes: a real win, with caveats

On labor reform, Sabharwal is genuinely upbeat. The labor codes removed about 70% of criminal provisions, 70-80% of forms, 50% of licenses. The threshold above which firms need government permission to retrench has gone from 100 to 300 employees — and UP is considering 50,000 or 100,000, which would effectively kill the provision. More importantly, much of the action has devolved to states. In Sabharwal’s framing, “29 chief ministers matter more than one prime minister for the next 25 years” because the daily life of an employer is local — labor markets, land markets, law and order, education, healthcare.

The China question, and the urbanization paradox

Asked whether India can do a China-style structural transformation without first solving land, Rajagopalan offers two paths. One: real fiscal federalism — push decisions on factories and conversion down to the gram panchayat level, the way China did at the township and village level. Diffuse the rents across tens of thousands of small jurisdictions and they become tractable. Two: leapfrog — wait for AI and high-yield tech to make agricultural productivity less land-dependent, the way mobile phones leapfrogged landlines.

On urbanization, she pushes back on the optimistic framing. Indians don’t move cleanly to cities precisely because of these laws. The eldest son moves; everyone else stays to occupy the unsellable, unconvertible, low-productivity plot. Removing the land laws would speed up urbanization both because land becomes liquid and because labor is willing to migrate permanently rather than seasonally.

Key Takeaways

Regulatory cholesterol, quantified: 67,000 compliances, 26,000 jailable, 6,700 filings.
Three-D agenda: decriminalize, deregulate, digitize. Decriminalization is furthest along; Jan Vishwas 2 + labor codes + Companies Act remove ~12,500 compliances.
Six pathologies behind the cholesterol — chiefly the compliance blind spot, instrument proliferation, and “prohibited till permitted.”
Land ceiling laws drive plot sizes to 1.08 hectares average, 0.18 in Kerala — directly causing India’s cereal yields to lag Bangladesh and Sri Lanka.
Land conversion laws impose an unjustifiable bureaucratic obstacle course with no externality basis — Rajagopalan recommends wholesale repeal, AP-style.
Labor codes removed ~70% of criminal provisions and devolved real power to states. Threshold for government permission to retrench up from 100 to 300; UP considering effectively eliminating it.
Federalism is the lever: 29 CMs matter more than one PM for the next 25 years, because the binding constraints are local.
Civil service reform is the missing piece — without it, the flow of new compliance will refill the stock.
Urbanization paradox: restrictive land laws slow urbanization rather than easing pressure on land regulation; the causal arrow runs the other way.

Claude’s Take

This is an unusually honest hour. Two people who have actually read the laws — Sabharwal’s team built a database of 67,000 compliances, Rajagopalan’s team read 50 statutes across 20 states — talking about specifics most policy panels gloss over. The phrase “regulatory cholesterol” is sticky for a reason: it captures that the harm isn’t a single bad rule, it’s the cumulative arterial blockage.

The strongest part is Rajagopalan’s land conversion analysis, where her training shows. The standard law-and-economics move is to weigh costs against benefits. She did the homework and found there is no benefit side of the ledger — every externality the conversion regime claims to address is actually addressed downstream at industrial permitting. That’s a much sharper finding than “the rules are too complex.” It’s “the rules have no defensible purpose.”

The honest weakness Sabharwal himself flags: the three-D approach is a stock cleanup, not a flow fix. Without civil service reform, the bureaucracy will simply generate new instruments to replace the ones being removed — and Sabharwal explicitly says civil service reform’s time has not come. So the bull case is: clear enough stock that growth picks up enough that the flow becomes politically harder to defend. The bear case: India runs the same loop it has run since 1991, where every reform wave eventually exhausts itself before the structural transformation completes.

The 1991-China comparison Ashtosh raises in Q&A — same per-capita GDP then, China is now 5x — is the gut punch that should anchor any conversation about Indian reform. Rajagopalan’s answer (decentralize like China, or leapfrog with technology) is plausible but speculative. The honest answer is nobody has solved this.

Score 8/10. Not 9 because there is some preaching-to-the-choir energy and the GW conference format limits real disagreement; not lower because the specifics — actual statute numbers, actual state comparisons, actual jail-provision counts — are the kind of detail you almost never get in this genre.