In July 2023, Indian state-run banks announced a target: $30 billion in inflows from a special overseas deposit scheme (FCNR(B)). By mid-July, they had mobilised roughly $10 billion. A 33% hit rate in two weeks sounds impressive — until you peel back the layers. The scheme is a textbook example of central bank liquidity engineering, but with mechanics that mirror the most fragile DeFi yield farms: high promised yields, a single point of trust (the government), and a maturity mismatch that could trigger a liquidity crisis when the deposits roll off.
Over the past six months, I have audited four rollup projects and dissected the incentive structures behind Aave’s interest rate models. The FCNR(B) scheme feels eerily familiar. The Reserve Bank of India (RBI) is offering non-resident Indians (NRIs) a guaranteed premium over global dollar rates, effectively subsidised by the Indian taxpayer. The banks swap the dollars for rupees at the RBI, inflating the central bank’s foreign reserves. On paper, it is a win-win: India gets a buffer against capital flight, NRIs get a safe, high-yield parking spot.
But code is law, and balance sheets do not lie. Let me show you why this $30 billion target is a revolutionary piece of financial theatre — and why the real vulnerability lies in the trust assumptions that nobody is auditing.
Context: The FCNR(B) Playbook
The Foreign Currency Non-Resident (Bank) scheme is not new. It allows NRIs to open fixed deposits in foreign currency (USD, GBP, EUR) with Indian banks, earning interest tied to international benchmarks (e.g., SOFR plus a spread). The key twist: the RBI allows banks to swap these dollars at a preferential rate, effectively eliminating forex risk for the bank while the central bank absorbs the exposure. In return, the RBI uses the dollars to pad its reserves, which stood at roughly $570 billion as of July.
The scheme’s allure is the yield. NRIs can earn 100–150 basis points above what they would get from a US Treasury or a global money market fund — all with an implicit government guarantee. For a country running a current account deficit of 2–3% of GDP, every dollar counts. But the structure creates a systemic risk interconnectivity that most analysts overlook.
Core: The Code-Level Analysis of a Central Bank Yield Farm
Let me break down the actual mechanics using the frameworks I apply to DeFi protocols. I call this the “liquidity bootstrap” — a temporary injection of external liabilities to mask structural weakness.
First, the interest rate model. Unlike Aave’s model, which adjusts supply and borrow rates dynamically based on utilisation, the FCNR(B) rates are arbitrary. The RBI announces a fixed spread over SOFR, and banks compete. There is no database link between the supply of NRI dollars and the demand for rupee liquidity. This is a centrally planned price — exactly the kind of rigidity I flagged in my 2020 audit of Compound’s governance model. When rates do not reflect market conditions, they create misallocations.
Second, the maturity transformation. The deposits are typically for 1–3 years. The banks lend these dollars (via swaps) to the RBI, which holds them indefinitely as reserves. But the banks still need to repay the principal and interest at maturity. If the external environment tightens — say, the US Federal Reserve hikes rates further — NRIs may not renew. The result: a sudden withdrawal that forces the RBI to either drain its reserves (selling dollars, weakening the rupee) or impose capital controls. This is the same “bank run” dynamic I described in my forensic report on Terra/Luna. The difference is that Terra had an algorithmic seigniorage model; India has a taxpayer-funded one.
Third, the leverage on bank balance sheets. State-run banks like State Bank of India (SBI) are the primary mobilisers. Their capital adequacy ratios are already under pressure from non-performing assets. Adding $30 billion in foreign deposits increases their liability side without a corresponding improvement in asset quality. The deposits are matched by RBI swaps, but those swaps are not risk-free — they depend on the RBI’s ability to honour the forex commitment. If the rupee depreciates 10%, the RBI’s swap liability in rupees increases, effectively transferring wealth from taxpayers to the banks.
Quantitative Rigor
Let’s run the numbers. Assume $30 billion flows in at an average rate of SOFR + 150 basis points. SOFR is currently 5.3%, so the cost to the banks is 6.8% per annum. The banks swap these dollars with RBI at the prevailing spot rate, receiving rupees. They then deploy those rupees into government securities yielding 7.0% or lend at 8–9%. The net spread is thin — maybe 50–100 basis points. Now factor in the RBI’s forex hedging cost (implicit, but real). The total subsidy is roughly $1.5–2 billion per year, paid by the RBI’s profits (which ultimately belong to the government). That is a revolutionary amount of money for a program that only shifts maturity, not solves the underlying trade deficit.
Contrarian: The Blind Spots Everyone Misses
Most headlines celebrate $30 billion as a sign of confidence. I see it differently. The scheme is a canary in the coal mine for three unspoken vulnerabilities.
Blind Spot #1: Trust in State-Run Banks
The mobilisation is concentrated in public sector banks. These banks have a history of governance issues, as I saw during my 2018 Solidity audit of a token contract that failed because the team ignored basic checks. NRIs are sending dollars to institutions that, in some cases, have negative net worth when mark-to-market. The implicit government guarantee is the only thing holding the system together. If that guarantee is even slightly questioned — say, due to a fiscal shock — the outflow could be instantaneous. DeFi protocols learned this the hard way: unbacked promises are not code.
Blind Spot #2: The Maturity Cliff
The deposits are mostly 1–3 year tenors. In 2024–2025, India will face a balloon of maturing FCNR(B) deposits. If global liquidity is still tight, the rollover rate could collapse. The RBI will then have two choices: drain its reserves (weakening the rupee) or offer even higher rates (raising the subsidy). Either way, the cost escalates. This is the same “death spiral” dynamic I identified in Terra’s bond mechanism: a self-referential dependence on continuous inflows.
Blind Spot #3: The Double-Counting of Reserves
When the RBI receives $10 billion via swaps, it records them as foreign exchange reserves. But those dollars are not “free” — they are borrowed and must be repaid with interest. Economists call this “reserve adequacy” but it is actually “reserve leverage.” The true free reserves (net of swap liabilities) are much lower. During my research on Layer 2 ZK-rollups, I saw a parallel: protocols that claimed high TVL but included their own tokens as collateral. The RBI’s reserves are similarly inflated by swap obligations.
Takeaway: The Vulnerability Forecast
The FCNR(B) scheme is a brilliant short-term fix — but it is not a sustainable solution. It is a yield farm that delays the inevitable. India’s structural current account deficit will not disappear because NRIs park their savings for two years. The real question: what happens when the farm closes?
The next crisis may not come from DeFi; it may come from a central bank that borrowed too much from its own diaspora. As I wrote after the Terra collapse, “code is law until it is not.” In this case, the code is the contract between the RBI and the NRIs. And that contract has a hidden clause: trust in the state.