A central bank that dislikes where its currency is can do three things, and they differ in cost and in evidence of effectiveness.
Sterilised vs unsterilised intervention
Unsterilised intervention means buying or selling foreign currency and letting the domestic money supply change as a result. Buying dollars injects domestic currency; selling dollars withdraws it. This works, because it is monetary policy, and the exchange-rate effect comes through the same channel as any other change in the money stock. The price is that the bank has surrendered its domestic monetary target to its exchange-rate target, which is the trilemma stated as a balance-sheet operation.
Sterilised intervention means buying foreign currency and simultaneously draining the domestic currency injected, so the money supply is unchanged. Only the composition of assets held by the private sector changes, which is why it can work at all only in the portfolio-balance world where domestic and foreign bonds are imperfect substitutes.
Advertisement
What does sterilised intervention cost?
Sterilisation has a running cost and the arithmetic is simple enough that no bank is unaware of it. Suppose a central bank buys $5 billion at ₹84.5000, injecting ₹42,250 crore. It earns the foreign reserve yield, say 4.35%, on the dollars it now holds, and pays the domestic rate, say 6.60%, on the bills it issues to drain the rupees. The carry is −2.25 percentage points, so the operation costs:
₹42,250 crore × 0.0225 = ₹950.625 crore a year
which is about $112.5 million at the same rate, or ₹190.125 crore for each billion dollars of intervention held for a year. Left unsterilised, that same purchase would have expanded a ₹45 lakh crore monetary base by 0.938889%. Neither number is large on its own. Both compound if the bank intervenes every month for a decade, and the accumulated sterilisation cost is a real fiscal item that shows up in the central bank's transfer to the government.
Does foreign exchange intervention work?
The evidence on whether intervention works splits cleanly by market type, and the split is worth stating because it is routinely ignored in both directions.
In deep, floating developed markets, sterilised intervention has small and short-lived effects on the level of the rate. It moves volatility and it can move the rate for hours or days by signalling policy intent, and studies of coordinated interventions find larger effects than unilateral ones. What it does not reliably do is move a floating major currency's level against a determined market, because the flows required dwarf any central bank's balance sheet.
In emerging markets the picture is different, and the reason is mechanical rather than mysterious. Markets are shallower, so a given intervention is a larger share of turnover. Capital controls limit how much offsetting flow can arrive. And reserves are frequently large relative to the domestic market. Sterilised intervention in those conditions has measurable effects on both level and volatility, and central banks that intervene routinely, including the Reserve Bank of India, are operating in a setting where the tool has demonstrated power.
What are capital controls, and how does India use them?
Capital controls are the third instrument, and the profession's view of them has moved. The pre-2008 consensus treated them as a distortion to be dismantled. The IMF's own institutional view since 2012 accepts capital-flow management measures as legitimate in specific circumstances: a surge of inflows that cannot be absorbed, an outflow crisis, or a situation where the alternative macroeconomic tools are exhausted. What the evidence supports is narrow. Controls on inflows, particularly measures that lengthen the maturity of what arrives, have some record of shifting the composition of flows toward the sticky end. Controls on outflows, imposed during a crisis, have a much worse record and a reputational cost that outlasts the crisis, because investors remember which countries closed the door.
India's framework is the working example for readers here. It has four parts. A managed float, in which the central bank smooths rather than targets. A legislated framework governing current and capital account transactions. A liberalised remittance scheme carrying a per-person annual limit for resident individuals. And an external commercial borrowing framework setting who may borrow abroad, at what cost ceiling and for what end use. Verify the remittance limit against the Reserve Bank's master direction, which is amended by circular. Every one of those numbers moves that way, so read the primary source and record the date you read it.