The Automatic Feedback Loop—and Its Friction
The fixed exchange-rate arrangement now has a clear mechanical channel: FX transactions change bank reserves, reserves affect the domestic money-market rate, and the rate can change capital flows. This chapter asks when that feedback is smooth and when it breaks down. The reader’s practical task is to distinguish a temporary reserve adjustment from a friction that transfers pressure into bank funding, market depth, and financial stability—even while the exchange rate itself remains fixed.
When automatic adjustment should work
The simplest version of the peg assumes that banks can move freely between local and foreign currency, that foreign-exchange transactions are inexpensive, and that money can settle quickly across markets. It also assumes that banks can reallocate reserves through an active interbank market. Under those conditions, a small reserve shortage raises the domestic short-term rate, the rate difference attracts capital inflows, and the inflow supplies foreign currency and reserves. A small surplus lowers the rate, encourages outflows, and reduces the surplus.
This mechanism is not a promise that the domestic rate will equal the anchor-country rate at every moment. It is a claim about direction and feedback. The local rate should move toward the level consistent with the fixed exchange rate, adjusted for the risk and transaction costs of holding local assets. The rate itself becomes the signal that tells investors and banks whether local liquidity is scarce or abundant.
Three conditions are doing most of the work.
First, FX flows must respond to the rate signal. If a higher local rate does not attract foreign funds because investors doubt the country’s institutions or fear a future devaluation, the rate can rise without replenishing reserves. If a lower rate does not send excess funds abroad, a reserve surplus can persist. The direction of the theoretical response may be intact while the magnitude is too small to close the gap.
Second, the response must arrive in time. A payment may be due today while the offsetting capital flow settles later. Time-zone differences, settlement conventions, or the need to aggregate several counterparties’ transactions can leave a bank short even if the market would eventually correct the shortage. A delayed response is not the same as no response, but a delay can make an otherwise manageable shock expensive.
Third, the domestic interbank market must redistribute reserves effectively. One bank may be short while another is long, but the surplus bank must be willing and able to lend. Counterparty limits, credit concerns, or a shallow market can stop the redistribution. In that case, the rate may overreact to a shortage or barely respond to a surplus. The reserve signal is then distorted by the market through which it is supposed to travel.
These assumptions also explain why a fixed rate does not eliminate a risk premium. A foreign lender cares about the anchor currency, but also about the local banking system, the fiscal authority, the depth of the domestic market, and the credibility of the central bank’s liquidity management. The IMF framework describes this country-specific premium as capable of changing during stress. The local rate can therefore sit above or below the foreign rate without immediately indicating that the peg has failed.
A useful diagnostic starts by separating the symptom from the cause. The symptom may be a jump in the overnight rate. The cause may be a reserve shortage caused by an FX outflow, a government payment, or a rise in precautionary reserve demand. The contributing condition may be weak interbank credit or a high country-risk premium. The consequence may be higher bank funding costs, reduced lending, or a speculative response. Looking only at the rate hides those distinctions.
How friction turns a shock into stress
Transaction costs create the first wedge. If buying and selling foreign currency involves a spread or other financing cost, a small difference between the local and foreign rate is not enough to trigger a flow. Capital moves only after the expected return exceeds the cost of moving and funding it. The domestic rate can therefore fluctuate inside a band around the anchor-consistent level without producing an immediate FX response.
Country and counterparty risk create a second wedge. A local bank may pay more to borrow than a foreign bank would pay for the same currency because lenders are pricing the risk of the country, the bank, or the collateral. In that setting, the relevant comparison is not just the local policy rate against the anchor rate. It is the return after the lender has accounted for the possibility of delayed settlement, default, restrictions, or a future change in the exchange-rate commitment.
Market depth creates a third. A well-developed interbank market can absorb a reserve shortage through many transactions among banks. A shallow market can turn the same aggregate movement into an abrupt price change because a small number of institutions must compete for funds. In a surplus, the problem reverses: abundant reserves can reduce the need to trade, making the market less practiced and less informative. The immediate symptom is low or unstable turnover; the deeper consequence is weaker capacity to absorb a later shock.
The IMF’s terms-of-trade example shows how these wedges interact. Suppose an adverse shock reduces the supply of foreign currency. At the initial rate, demand for foreign currency now exceeds supply. Banks or the central bank must meet the excess demand by selling foreign currency, which reduces foreign reserves and the banking system’s reserve supply. With fewer reserves, the domestic money-market rate rises. That higher rate may attract capital inflows and gradually offset the original shock.
The word “gradually” matters. If the inflow is large and fast enough, the rate eventually returns toward the level consistent with the anchor currency. If it is delayed or too small, the local rate remains high. Banks pay more for funding, borrowers face tighter conditions, and institutions that cannot refinance easily may become illiquid. The exchange rate can remain at its promised value throughout this sequence. Defending the peg has shifted the adjustment burden into domestic rates and bank balance sheets.
The reverse sequence is not harmless either. An autonomous factor can inject reserves, or foreign-currency inflows can leave banks with more liquidity than they want. The domestic rate falls. If capital outflows do not respond, excess reserves persist. Banks may rely less on interbank funding and hold fewer incentives to develop the market. A system that looks calm because rates are low can become more fragile if the market loses the activity needed to price and redistribute liquidity.
The direction of the initial shock is not the only thing that matters. The IMF framework also considers a rise in banks’ precautionary demand for reserves. Suppose banks become more concerned about settlement or counterparty risk while the supply of reserves stays unchanged. Their demand for reserves shifts outward. The new scarcity raises the domestic rate, which can induce banks to convert foreign currency into local currency through the central bank. That inflow increases the reserve supply and can bring the rate back toward its earlier level. In a liquid, trusted market, the adjustment may be short-lived. In a market where every bank is trying to protect itself, the precautionary demand can be larger and more persistent, making the same reserve supply inadequate.
This case is different from an adverse terms-of-trade shock. In the terms-of-trade case, the first disturbance is a reduced supply of foreign currency. In the precautionary-demand case, banks want more reserves even though the existing supply has not changed. The observed symptom—a higher money-market rate—can look similar, but the underlying cause is different. That distinction matters because a central bank trying to interpret the rate needs to know whether it is responding to a foreign-exchange drain, a domestic balance-sheet movement, or a change in banks’ willingness to hold liquidity.
The model also allows a positive liquidity shock, such as government spending that reduces the government’s account at the central bank. The banking system receives reserves, the domestic rate falls, and demand for foreign currency may rise if the local return becomes too low relative to the anchor. The central bank can then sell foreign currency, withdrawing the excess reserves and restoring the prior money-market position. If that FX response is delayed or if capital outflows are weak, the reserve surplus remains in the system. What first looked like a benign injection can become a persistent operating condition with consequences for trading and future reversals.
The common thread is that a peg must manage both reserve supply and reserve demand. A central bank can face stress even when foreign reserves are not falling if banks suddenly want to hold more domestic liquidity. It can face market erosion even when the banking system is flush with reserves. Reading only the FX side or only the rate side will miss half of the mechanism.
This creates two different failure paths:
- Persistent shortage: high rates, expensive funding, precautionary reserve hoarding, and possible illiquidity or solvency pressure.
- Persistent surplus: low trading activity, weaker market development, and a larger pool of liquidity that may complicate a future reversal.
Neither path proves that the peg is unsustainable. It shows that exchange-rate stability and financial stability are different outcomes. A central bank can keep the external price fixed while the banking system absorbs increasingly severe internal volatility.
The state diagram is deliberately not a prediction of a single crisis. It describes how the same fixed-rate commitment can pass through different operating states. A friction is a condition that makes the first adjustment slower or more expensive. A persistent shock is one that outlasts the market’s ordinary response. A shortage or surplus is a reserve outcome. Financial stress and market erosion are consequences that can occur without an immediate movement in the exchange rate.
A useful way to test the diagnosis is to ask whether the system’s response is self-limiting or self-reinforcing. A self-limiting response moves the rate enough to attract or repel flows, restores a workable reserve position, and then fades as the original imbalance is corrected. A self-reinforcing response raises precautionary demand, increases the risk premium, or weakens the market so that the next reserve movement produces an even larger rate response. The first can look volatile in the moment but settle without lasting damage. The second can turn a small initial disturbance into a persistent funding problem.
Timing is part of that test. A rate increase that lasts for a few settlement periods may be the cost of bringing a flow through. A rate increase that persists because every bank is hoarding reserves points to a different problem. Similarly, a low rate during a reserve surplus can be a temporary sign that the market has ample liquidity, or it can be a sign that interbank trading has become unnecessary and the market’s shock-absorbing function is deteriorating. The level alone is not enough; the duration, counterparties, and reserve behavior around it matter.
That distinction helps eliminate several tempting but incomplete diagnoses. If the overnight rate jumps, the cause may not be a deliberate tightening of the policy stance; it may be an unoffset reserve drain. If the rate stays low, that does not prove that liquidity conditions are healthy; a structural surplus may be suppressing the rate while reducing interbank activity. If capital flows eventually reverse a shock, that does not prove the process was costless; banks may have carried high funding costs while waiting for the flow to arrive.
The diagnostic question is therefore not simply whether the rate moved in the “right” direction. It is whether the movement was proportionate, timely, and transmitted through markets that can continue functioning. A rate increase that attracts enough capital to restore reserves may be part of a healthy feedback loop. The same rate increase, in a market with high risk and weak counterparties, may become a funding shock. A reserve surplus that lowers rates can support activity in one setting and hollow out the interbank market in another.
The framework also places a boundary around what can be concluded. These are theoretical mechanisms derived from a stylized model of fixed exchange rates with an open capital account. They identify conditions under which instability becomes more likely; they do not establish that every peg experiences every failure path or that a particular country is currently in one of them. Actual outcomes depend on market institutions, risk perceptions, settlement arrangements, fiscal flows, and the central bank’s available tools.
The central conclusion is sharper than “frictions matter.” A peg’s automatic adjustment mechanism is itself a source of exposure. When the loop works, it can absorb ordinary flows without large discretionary intervention. When it works slowly or unevenly, the same loop can make rates, reserves, and bank liquidity carry the burden of defending the exchange-rate commitment. The next task is to combine these mechanisms into a stress test: identify where a shock enters, where it is absorbed, and where it is merely displaced.