One kilogram of rice, six government prices

India does not have a price of grain. It has a ladder of administered prices, each one a policy lever, and the gaps between the rungs are where every subsidy, every incentive, and every distortion in the system lives.

Three features of this ladder define Indian farm pricing:

• The buy rung is a promise, not a market. MSP is announced for 23+ crops each season on CACP cost formulas (A2+FL +50%), but it binds only where a procurement machine stands behind it — which means purchase centres, mandis, and godowns, not gazette notifications.
• The sell rungs are policy dials. OMSS reserve prices, the ethanol rate, and issue prices move by executive order — sometimes twice a season (rice-for-ethanol: ₹2,800 → ₹2,250 in one January). Government adjusts the disposal price to steer private demand the way a central bank adjusts repo.
• Parallel ladders exist per crop. Sugarcane runs the same architecture in mirror image: a statutory floor mills must pay (FRP ₹355/qtl, +29% since 2019) against a frozen output price (sugar MSP ₹31/kg since Feb 2019) — a policy-made squeeze that produced ₹16,087 crore of farmer arrears by February 2026.

What the farmer actually receives

The distance between announced price and received price is the central empirical fact of the system. Three measurements, all from verified official series:

Coverage — who sells at MSP at all

Among households that actually sold the crop, 13% of paddy sellers and 16% of wheat sellers sold to a procurement agency (NSS 70th round unit data, 2012-13; the famous "6%" figure divides by all agricultural households). Official season counts: ~119 lakh paddy farmers paid in KMS 2024-25; 25.1 lakh wheat farmers in RMS 2025-26. For the other twenty-one MSP crops, coverage is a few percent in the best year and zero in most.

Volume — the share of the harvest the promise reaches

Reliability — the variance that breaks sowing decisions

Year-to-year coefficient of variation of procured volume: rice 24%, wheat 21% — versus gram 80%, tur 109%, bajra 110%, maize 120%, urad 143%, soybean 163%. Cotton is the pure case: 124 lakh bales bought in 2019-20, zero in 2021-22 and 2022-23, ~100 lakh again in 2024-25. For a farmer choosing what to sow, a guarantee with 120% variance is not a guarantee; it is a lottery whose ticket is a season's income.

The trade boundary: export and import parity

Domestic administered prices float inside a boundary set by world prices. The rules of thumb per crop class:

• Rice — exports clear the market, not the government's stock. DGCI&S unit values put non-basmati exports at −12% to +2% of the MSP-derived cost of rice across FY24–26 (FY26: record 15.0 MMT at ~$390/t after the Sep–Oct 2024 ban unwind; basmati earns +100–140%). Private exporters at market paddy prices clear ~40% of world trade — but the state's costlier procured stock cannot exit at these unit values without loss, which is why its disposal runs through discounted ethanol (₹2,320) and OMSS instead. The 2022–24 bans were the same boundary managed in the other direction.
• Maize — export parity below MSP, ethanol above it. Indian maize FOB is ~$220–245/t ≈ ₹1,890–2,110/qtl against a ₹2,400 MSP. Exports cannot absorb an MSV purchase; the ethanol loop (NAFED delivery to distilleries at ₹2,291/qtl, blending mandate demand) is the channel that can — and already does.
• Pulses — the parity sign just flipped hard. DGCI&S FY26: tur lands at CIF ₹5,659/qtl — 29% below its ₹8,000 MSP (from +10% above in FY25); masur −23%, chana −15%, yellow peas −44% vs gram MSP. Tariffs are visibly chasing the gap (chana and masur to 10%, yellow peas to 30%) while tur/urad stay duty-free to Mar-2026 under the Mozambique–Myanmar–Malawi pacts. Pulses MSV without tariff coordination is fiscally endless — duty policy and the volume guarantee are one instrument, not two. Edible oils are the extreme case: FY26 CIF undercuts MSP-linked crush economics by 30–40% even after duty, on a $19.5 bn import bill.
• Wheat — the marginal case. Export-viable only in world-price spikes (2021-22: record procurement and record exports, then the May 2022 ban a fortnight after a heatwave). Both boundaries are policy-active.

The MSP + MSV ratio: one number, tuned per crop

Define θ (the MSV ratio): the share of a crop's marketed surplus the government commits — in advance, before sowing — to purchase at MSP. Today θ is an accident of infrastructure: ~0.35 for rice, ~0.3 for wheat, ~0 for everything else. The reform is to make θ an explicit, published policy variable.

Why a partial ratio creates outsized farmer value

Variance falls with the square of the ratio: a θ of just 0.3 removes 51% of price-revenue variance; θ = 0.5 removes 75%. The first tranche of guaranteed volume does most of the stabilization work — because what destroys farm incomes is the crash-year tail, and a pre-committed volume floor is exactly a put option on that tail. Three further farmer-side effects compound it:

• Credit: a government purchase commitment is bankable collateral; kisan-credit pricing against an assured θ·MSP cash flow beats pricing against a lottery.
• Sowing signal: announced before sowing, per-district volume caps steer acreage the way Haryana's ₹7,000/acre diversion payments have failed to — because they change expected revenue, not just add a side payment.
• Bargaining floor: even the (1−θ) share sold privately clears nearer MSP when traders know the state stands ready at the margin — the documented mandi-price support effect of procurement presence.

Why a capped ratio creates government value

• The exposure is capped and known at budget time — unlike the sugarcane FRP model, where an uncapped statutory obligation with no viability check manufactured ₹16,087 crore of arrears, or uncapped bajra buying that ended in loss-making resale.
• The purchased volume is a merchant book, not a write-off. The state already monetizes stock through four proven channels: OMSS e-auctions to processors (71+ LMT wheat in 2023-24), the ethanol programme (52–72 LMT rice; a tender clause now mandates 40% FCI-grain sourcing for distillery supplies), Bharat-brand retail, and PDS/welfare offtake. MSV extends a working disposal machine to new crops — pulses to dal mills, maize to feed and distilleries, millets to PDS/ICDS mandates.
• Storage physics is respected. Public capacity (~777 LMT covered) caps feasible procurement at roughly 30% of the ~261 MT marketed surplus of the major crops. Basis: 776.66 LMT is conventional covered capacity — FCI-owned plus hired (CWC, SWC, PEG, PWS-2010, private, state government) plus capacity held by state agencies — from the DFPD Foodgrain Bulletin, November 2024. It deliberately excludes the 64.62 LMT modern silo programme (hub-and-spoke Phase-I awarded, rail/road-fed and circuit-model silos, and units under construction), which is a later vintage (PIB, 31 December 2025) and not yet fully operational. Adding it gives the 841 LMT all-agency figure used in the state-map piece; the 917.83 LMT figure quoted elsewhere is FCI+State covered plus CAP (open-plinth) storage at 1 July 2025, and open plinth is not usable for the multi-season buffers modelled here. The three numbers differ by scope, not by disagreement. θ per crop is chosen inside that envelope — the constraint becomes a design input instead of a crisis.
• Counter-cyclical release earns the spread. Buy at MSP in the glut, release through OMSS in the tight year: the state's buffer function, done deliberately, partially self-finances.

The grain book: what the state would buy, and where it would sell

Priced against the market as it stands today — every mandi report in the country on 23 July 2026 (17,247 Agmarknet records), medians per crop against 2025-26 MSPs:

Read the highlighted rows: the grains MSV would actually buy today are the millets (bajra, ragi, jowar), maize, tur, moong, and groundnut — the crops where a majority of mandi reports sit below MSP right now. Mustard, soybean, urad, gram, and cotton trade above MSP today: their θ costs nothing while dormant. That is the self-limiting property of the design — the volume guarantee spends only where and when distress exists, and every buy-row has a named sell-channel at a stated price before the first quintal moves.

Three design rules the evidence keeps repeating

1. Announce θ·Q before sowing; never after. Post-harvest volume decisions are politics; pre-sowing volume commitments are policy.
2. Contract the exit before the entry. Every quintal procured under MSV has a named disposal channel — a standing e-auction calendar, a blending-mandate allocation, a PDS quota — priced and published. The July 2023 ethanol pause showed what happens to processor trust when the exit is discretionary.
3. Budget the rung-gap openly. Selling at ₹2,320 what costs ₹4,000 is a legitimate policy transfer — when it appears as a line item, not as FCI balance-sheet sediment.

Making the "put option" language precise

This piece has already been describing MSV informally as a floor the farmer holds against a crash-year price — which is, in the formal language of insurance and options economics, exactly what it is. Kenneth Arrow's work on the economics of uncertainty and insurance treats any instrument that pays out specifically in the bad state of the world, and nothing extra in the good state, as insurance against that state; a farmer who sells at market price P_mkt whenever P_mkt exceeds MSV, and sells to the government at MSV whenever it doesn't, holds a position identical in payoff structure to owning a put option struck at MSV on their own harvest — capped downside, uncapped upside, in exchange for a "premium" that here is paid not in cash but in the government's own carrying cost of the procurement-and-disposal apparatus. The variance formula already used above, Var(R) = (1−θ)²·Var(P_mkt), is precisely the variance-reduction identity for a portfolio holding a fraction θ hedged at a fixed strike — the same algebra a derivatives desk would use to price the government's implicit liability.

Ronald Coase's 1937 and 1960 work on transaction costs and contract design supplies the second half of the picture: Coase's core insight is that outcomes depend less on who holds a formal entitlement than on whether the contract around it is complete — whether every contingency has a named, priced path, so no party is left renegotiating under duress after the fact. The design rule directly above this paragraph, "contract the exit before the entry," is a Coasean rule in practice: the 2023 ethanol pause was costly specifically because the exit channel was left open-ended rather than contracted, so processors bore a renegotiation risk that a complete contract would have priced in from the start.

This is a formal restatement of the piece's own already-informal "put option" framing, not a claim that Indian procurement agencies model MSV as a derivative internally.

The value split, stated plainly

For the farmer, MSP+MSV converts an announced price into a bankable quantity: three-quarters of revenue variance removed at θ = 0.5, credit access against a sovereign cash flow, and a sowing-time signal that actually moves acreage.

For the government, it converts an open-ended moral commitment into a budgeted, storage-feasible position with four working monetization channels — at roughly a third of the fiscal cost of a universal legal MSP, with the crash-year tail (the part that causes farm distress and its politics) covered first.

The one-line design: MSP without volume is a promise; MSV without disposal is a warehouse fire in slow motion. The instrument that works is a pre-announced volume ratio with a pre-sold exit.