Hook
The Reserve Bank of India's FCNR(B) deposit scheme has mobilized nearly $10B in weeks, targeting a total of $30B. At the protocol level, this looks less like a monetary policy tool and more like a yield farming campaign with a 3-year lockup. The code is simple: deposit USD at a premium rate, get INR-denominated returns, and hope the exit doesn't trigger a bank run. Tracing the gas leaks in the 2017 ICO ghost chain taught me to smell state-backed incentives before market narratives solidify.
Context
FCNR(B) stands for Foreign Currency Non-Resident (Banking) account. It allows non-resident Indians to park foreign currency in Indian banks for terms of 1-3 years, earning interest rates linked to LIBOR or SOFR plus a spread. The RBI uses these dollars to bolster foreign exchange reserves and defend the rupee. State banks, which dominate the scheme, act as liquidity providers. The incentive structure is reminiscent of a DeFi liquidity mining program: deposit assets, earn yield, but with a predetermined vesting schedule and no early withdrawal.
Core
From a technical standpoint, the FCNR(B) scheme is a permissioned liquidity pool. The RBI issues a governance proposal (policy change), banks deploy capital as validators, and NRIs provide liquidity. The yield is deterministic: fixed interest over 1-3 years. But the risk is asymmetric. Based on my audit of Anchor Protocol’s unsustainable yield mechanics in 2022, I see a similar pattern here: the spread between deposit costs and domestic lending yields is subsidized by the RBI’s forward commitment. The protocol’s solvency depends on future capital inflows to roll over maturing debts.
Quantifying the Match
Let’s run the numbers. Assume $30B is locked at a 4% average LIBOR-based rate. Over three years, the Indian government pays ~$3.6B in interest. The implied cost of preventing rupee depreciation? If $30B prevents a 5% drop in the INR against the USD (from 83 to 87), the RBI saves $150B in import cost exposure. But the TVL is concentrated in state banks like SBI, which lack the capital buffers of private DeFi protocols. The protocol’s liquidation mechanism is non-existent: no algorithmic market maker, no flash loan protection. If the dollar strengthens suddenly, the collateral (rupee reserves) devalues, creating a negative feedback loop.
Smart Contract Risks
During my 2020 deep dive into Uniswap V2, I modeled impermanent loss. For FCNR(B), the equivalent is currency risk: NRIs deposit USD but redeem in USD after conversion to INR. If the rupee depreciates > the interest earned, they suffer net loss. The code doesn’t hedge this. The RBI’s “oracle” is the spot forex market, which is manipulated by whales. The state banks’ execution layer has no slippage controls. Silicon whispers beneath the cryptographic surface: the real bug is the assumption of sovereign credit default guarantees. In 2026, when I audited a zero-knowledge proof system for AI inference, I saw how centralized verification leads to single points of failure. Here, the failure point is the Indian government’s credit rating.
Contrarian
The bullish narrative? $30B inflow strengthens reserves, stabilizes the rupee, and boosts market confidence. But the blind spot is the maturity wall. In three years, if global liquidity tightens, NRIs may not roll over deposits. This is a classic time-lock bomb — similar to Terra’s 21-day unbonding period. The protocol lacks a gradual unstaking mechanism. Patching the silence between protocol updates: the RBI should implement a linear vesting schedule or allow partial early withdrawals with a penalty fee (like a slash in Cosmos). Otherwise, the exit frenzy will replicate what I observed in the 2022 bear market: a forensics race where code remembers what auditors missed.
Takeaway
The FCNR(B) scheme is a sovereign DeFi experiment with smart contract bugs that haven’t been patched yet. The market will celebrate the TVL, but the real vulnerability is the absence of a redemption circuit breaker. Will the RBI add a pause function before the next rate hike?
— Decoding the chaos of the bear market ledger.