The Turkish central bank cut interest rates over 2021-2023, on the politically driven view that higher rates cause inflation. This column analyses the policies used instead: foreign exchange interventions, a foreign-exchange-protected deposit scheme, and eventually financial repression and soft capital controls. These interventions did not remove currency risk but moved it onto the government, widened sovereign spreads, and converted a monetary problem into a fiscal one. After the 2023 elections, orthodox policy returned. However, the costs of this experiment are large and still being paid, making the eventual adjustment larger than the one avoided.
Political pressure on central banks is building again, most visibly on the Federal Reserve (Drechsel 2024). What may such pressure cost an economy? Türkiye offers the clearest case study. Between 2021 and 2023 its central bank cut interest rates while central banks elsewhere were hiking (Figure 1), on the politically driven official view that high interest rates cause inflation rather than cure it. The outcome was the opposite of what the policy’s architects expected, and close to what standard theory predicts. Gürkaynak et al. (2023) document it: the lira depreciated sharply, inflation accelerated to 85%, and long-term interest rates rose rather than fell.
Figure 1 Cumulative change in policy rates since July 2021
Source: BIS.
But that is half the story. Countries in this position do not sit still. They look for alternatives — doctrines that justify low rates, and financial engineering that makes low rates survivable. A government that cannot raise rates still has to keep its currency from falling and its banks funded. So it turns to foreign exchange intervention and government guarantees of various kinds. The Bank of Japan’s US-backed yen intervention in July 2026 is a stark example. In recent work (Kara and Simsek 2025) we document the Turkish episode and model what these innovative policies actually achieve, and what they cost, when the interest rate cannot respond. Our answer is that the second act is where most of the damage occurs.
Türkiye’s Great Policy Experiment of 2021–2023 was politically motivated. Between 2019 and 2021, three central bank governors were dismissed after failing to deliver the rates the government wanted (Figure 2). The cuts began in September 2021, when inflation stood at 19% (Figure 3). Real interest rates turned deeply negative, credit accelerated, the current account deficit widened, and the lira slid.
Figure 2 Term length of Central Bank of the Republic of Türkiye governors since 1996
Note: *As of August 2026. Size of the bars shows the total years the respective governor served.
Source: Central Bank of the Republic of Türkiye (CBRT).
Figure 3 Türkiye’s policy rate and CPI inflation
Notes: Vertical lines mark the September 2021 cuts, the introduction of the FX-protected deposit scheme (KKM), the 2023 return to orthodoxy, and KKM’s closure.
Sources: CBRT, TURKSTAT.
The first line of defence was the foreign exchange market. Türkiye intervened heavily, partly through opaque arrangements in which public banks sold central bank reserves. Our model shows that when interest rates are held artificially low, intervention buys only temporary relief. It delays the adjustment, exhausts the reserve buffer, and leaves the country exposed to a sudden stop. Türkiye’s arrived in December 2021, when the dollar rose 80% against the lira in a single month.
Instead of raising rates, the government tried something new. Under the foreign-exchange-protected deposit scheme (kur korumalı mevduat, or KKM), a lira depositor was promised the higher of the lira interest rate or the depreciation of the lira against the dollar. Put differently, the government wrote depositors a free call option on the dollar, struck at the policy rate (Figure 4).
Figure 4 Foreign-exchange-protected deposit scheme (KKM) payoffs resemble a call option on the dollar, struck at the policy rate
Notes: If the policy rate is 10% and the lira depreciates 40%, the depositor receives 40% and the government pays the 30% difference.
The design addressed the immediate problem. Our model demonstrates that KKM raised the return savers expected to receive without raising the rate borrowers expected to pay, with the government absorbing the difference. It also eliminated the foreign exchange risk embedded in lira deposits. These predictions are consistent with what followed. The run out of the lira stopped, the lira stabilised (Figure 5), and dollarisation slowed. Take-up reached roughly $140 billion by mid-2023 — about 10% of GDP and a fifth of all deposits.
Figure 5 Nominal and real effective exchange rates
Notes: The dashed line marks KKM’s introduction. Sources: Yahoo Finance, BIS.
The difficulty lies in what the government had taken onto its books. The cost of the KKM guarantee rises with the depreciation of the lira — that is, in exactly the circumstances that made it necessary. And because the payoff is option-like, the cost rises more than proportionately with the exchange rate. Meanwhile, the central bank continued to borrow foreign exchange from banks and sell it to finance the external deficit. By mid-2023 the central bank’s net foreign exchange short position together with KKM liabilities had reached about $200 billion, over 15% of GDP (Figure 6, top panels).
Figure 6 Net foreign exchange position and fiscal exposure of the Turkish central bank
Notes: Clockwise from top left: KKM outstanding; the central bank’s net foreign exchange position including KKM; five-year credit default swaps (CDS), Türkiye versus emerging markets; government foreign exchange debt plus KKM, % of GDP. Sources: CBRT, Ministry of Treasury and Finance, Bloomberg.
KKM and FX interventions did not remove currency risk from the economy. They moved that risk onto the government. Our model shows that this creates a trap. Suppose depositors come to doubt that the government will honour the guarantee. They withdraw. Withdrawal depreciates the lira, depreciation raises the cost of the guarantee, and the higher cost validates the original doubt. The crisis is self-fulfilling, and it needs no external shock: because the guarantee was itself holding the currency up, leaving it is the shock. This also implies that a lower policy rate makes the trap more likely: the currency becomes weaker without the guarantee, so the shock of leaving the guarantee is larger.
Markets priced the risk of a debt crisis. Türkiye’s sovereign spreads widened relative to comparable emerging markets after KKM was introduced, even though the debt-to-GDP ratio initially remained moderate. That pattern fits markets assessing the contingent fiscal exposure and its interaction with the external balance rather than conventional concerns about debt sustainability (Figure 6, bottom panels). Depositors also seem to have doubted that the government would ultimately pay. KKM take-up stalled below a quarter of total deposits even though its option-like payoff dominated both ordinary lira and foreign exchange deposits.
This crisis never arrived in Türkiye, but the possibility of it shaped policy. Once such an outcome becomes possible, the expectation of depreciation can set it in motion, so the authorities cannot afford to let the lira move much and are trapped in a tightly managed foreign exchange regime. The exchange rate had stopped being merely a price. It had become a fiscal variable, and defending it became a budgetary and financial-stability necessity. Unable to use the policy rate or let the lira adjust, the authorities resorted to financial repression and soft capital controls. These included requirements for exporters to surrender their foreign currency earnings, credit quotas, reserve requirements tied to KKM targets, and restrictions on firms’ foreign exchange deposit holdings. The standard escape route was also closed off. A government with large local-currency debts can inflate its way out of them. A government that has written options on the dollar cannot, because printing lira depreciates the currency and raises the obligation faster than inflation erodes it. In substance, these were dollar liabilities.
After the 2023 elections a new economic team returned to orthodox policy and raised the policy rate from 8.5% to 50%. KKM was wound down and formally closed on 23 August 2025. The costs of the experiment are still being paid, and the visible ones are large. The central bank’s losses ballooned in 2023, driven by a KKM cost of $35 billion, or 3% of GDP. This outlay amounted to direct money printing. Three years into the disinflation programme, inflation stood at 32% in July 2026 against a 5% target, with the effective policy rate at 40%. Growth has run below potential for two years, and the lira’s sharp real appreciation has squeezed exporters and industrial production.
The less visible costs are the more instructive. Credibility, once spent, is slow to rebuild: the central bank had to keep interest rates excessively high to attract the inflows needed to rebuild reserves and wind down KKM. Türkiye paid international investors carry-trade returns above 25% in 2024 to keep them holding lira. Rebuilding the reserve buffer compounded the bill, because the dollars had to be bought with high-yielding lira liabilities (Fanelli and Straub 2021). And inflation itself became harder to remove. Several things are holding it up, including lingering political uncertainty, the central bank’s late response to inflation, and a fiscal stance that initially remained loose. Yet part of the explanation is the legacy of the experiment itself. In our model inflation is inertial, so the longer an economy is held in the low-rate regime, the more entrenched inflation becomes, and the more output must be given up to remove it. Creative policies of this kind do not merely postpone the adjustment. They enlarge it.
Türkiye’s scheme was unusual in scale and design but not in kind. Many emerging markets have used the government balance sheet to counter dollarisation induced by politically driven low rates. Mexico’s tesobonos, dollar-indexed peso bonds issued in 1994, first calmed markets and then amplified the collapse when the peso fell (Whitt 1996, Meza 2018). Brazil moved private currency risk onto the central bank through a special deposit facility in the 1970s and 1980s (Dalto 2019). And in 2015 Argentina’s central bank sold $17 billion of dollar futures at below-market rates rather than adjust the official rate. The cost, near 10% of the monetary base, fell to the incoming government (Sturzenegger 2019). In each case a government offered protection against depreciation rather than raise interest rates. The protection bought time, often until an election was safely past. Then the bill arrived. Türkiye avoided the collapse that Mexico suffered, but not the cost.
Under political pressure, schemes of this kind will always be tempting: they appear to solve the immediate problem without the political cost of a rate hike. Türkiye’s recent experience shows that they do not. A government that absorbs depreciation risk to avoid raising interest rates converts a monetary problem into a fiscal one. It buys time, yet it leaves the economy exposed to a crisis of confidence. Moreover, it makes the eventual adjustment larger than the one that was avoided. The political desire for low interest rates may culminate in a self-defeating process.
Source : VOXeu
