- Why Ghana’s Forex Auctions Struggled to Stabilise the Cedi Despite Billions of Dollars Sold
Foreign-exchange auctions can improve transparency and price discovery in Ghana’s currency market, but poorly calibrated supply targets and restrictions on pricing may instead intensify pressure on the cedi, according to a new Bank of Ghana working paper.
The study, prepared by Gershon K. Agbledzorwu of the central bank, examines Ghana’s three main foreign-exchange auction episodes from the introduction of the first auction in September 1986 to the suspension of regular forward auctions in October 2022.
Its central conclusion is uncomfortable for policymakers: an auction does not remove an underlying shortage of foreign exchange. Where the dollars offered are substantially below market demand, the auction may merely reveal the size of the shortage and push unsatisfied buyers into the wider market at higher exchange rates.
The most striking evidence comes from the seven wholesale spot auctions conducted between November 2016 and March 2017.
Banks submitted bids totalling US$694.50mn, while the Bank of Ghana sold US$248.75mn. This left US$445.75mn of demand unmet.
Each auction was accompanied by a sharp depreciation of the cedi on the day it was conducted, contrary to the expectation that the additional dollar supply would ease pressure on the currency.
“Each of the seven auctions was accompanied by sharp exchange rate depreciations, contrary to expectations,” the paper said.
“This was due to tight forex liquidity conditions in the market, evidenced by the high unmet demand on auction days.”
The finding challenges a common assumption in exchange-rate management: that selling foreign currency through a transparent auction is inherently stabilising.
Transparency can improve allocation, but it cannot compensate for an inadequate volume of foreign exchange or an exchange rate that does not reflect demand and supply.
The 2016-17 auctions were introduced partly to create a transparent mechanism for passing proceeds from Ghana’s annual cocoa syndicated loan to the banking system.
The programme was also intended to support market development and price formation as the Bank of Ghana reduced mandatory foreign-exchange surrender requirements for gold and residual cocoa proceeds.
A total target of US$180.50mn was announced across the seven auctions, but the central bank eventually sold US$248.75mn — US$68.25mn more than the intended amount.
Even this additional supply satisfied only about 36 per cent of the US$694.50mn submitted by banks.
The remaining demand did not disappear. Banks and businesses seeking dollars had to pursue foreign currency outside the auction, where the imbalance between supply and demand placed additional pressure on the cedi.
The paper found that the auctions were successful in terms of their operational mechanics. Allocations were generally conducted according to the best-price principle, beginning with the highest bidders.
Their macroeconomic outcome was different.
The central bank used reserves without achieving a lasting stabilisation of the exchange rate. Instead, the exercise produced “a gradual depletion of foreign exchange reserves, accompanied by sharp depreciations, especially on auction days”.
The study concludes that tight liquidity, unmet demand and the behaviour of participating banks indicated that the cedi was overvalued and required further adjustment.
This is an important distinction. An auction may be procedurally transparent while remaining economically ineffective.
The paper also identifies problems with the rules governing the prices banks could submit.
Section 6(d) of the 2016 auction guidelines restricted bids to a range within 2 per cent of the previous auction’s weighted average exchange rate.
The provision was intended to exclude extreme or outlier bids. But the study argues that it effectively became a form of price guidance, slowing the exchange rate’s adjustment towards the level required to clear the market.
Banks that urgently needed foreign exchange responded strategically. Some submitted bids based on the previous auction’s weighted average rate plus a margin approaching the permitted 2 per cent ceiling.
The paper described this as a “price war” among competing banks, with institutions seeking to improve their chances under the best-price allocation method.
The result was a contradiction at the heart of the auction. The system was intended to discover the market price of foreign exchange, yet the rules restricted how far participants could move away from the previously observed price.
“The 2 per cent cutoff margin in section 6d of the guideline implemented during the second episode implicitly served as a price guidance and contributed to slowing down the price adjustment process,” the study said.
Future auctions, the author recommends, should avoid price restrictions if their objective is genuine price discovery.
That does not mean a central bank must accept every speculative or clearly abnormal bid. It means that an auction cannot be expected to reveal the clearing price if participants are prevented from submitting rates that reflect their assessment of market conditions.
Ghana’s third auction experiment began in October 2019 and adopted an unusual deliverable-forward structure.
Instead of receiving dollars immediately, successful buyers paid the cedi equivalent upfront and received the foreign currency at an agreed future date.
Tenors included seven, 15, 30, 45, 60 and 75 days.
The arrangement was designed to reduce the practice of businesses buying and holding foreign currency ahead of future obligations because they were uncertain about its availability.
Companies with payments due in several weeks could bid for delivery at the appropriate maturity rather than competing for dollars in the spot market immediately.
The structure also offered the Bank of Ghana two advantages. It postponed the foreign-currency cash-flow effect until the delivery date and immediately removed the corresponding cedi liquidity from the banking system.
Between October 2019 and October 2022, the central bank conducted about 80 forward auctions and sold US$3.021bn.
Demand, however, reached US$9.169bn, leaving US$6.148bn unmet.
The average bid-cover ratio was 3.27, meaning banks requested more than three times the amount of foreign exchange available.
“On average, the required FX needed by banks is three times more than the available FX support,” the paper said.
The imbalance is even more notable because individual banks were prevented from submitting aggregate bids exceeding 20 per cent of each auction’s announced target. Demand might therefore have been higher if participants had not faced bidding limits.
The programme sold US$216mn more than its aggregate target of US$2.805bn, but still met only about one-third of the bids submitted.
Market feedback was broadly positive because the auctions provided additional liquidity and greater certainty about future dollar availability. Yet they could not eliminate the structural gap between foreign-exchange demand and the central bank’s capacity to supply it.
Most successful allocations were concentrated in shorter tenors.
Of the US$3.021bn sold, US$1.386bn went to seven-day contracts and US$959.25mn to 14- or 15-day contracts as reported in the study’s summary table. A further US$467mn was allocated at 30 days.
Only US$136.75mn went to the 45-day tenor, US$56mn to 60 days and US$16.50mn to 75 days.
The first three maturities therefore accounted for more than 93 per cent of the amount sold.
This concentration suggests that businesses and banks remained primarily concerned about near-term dollar availability rather than using the auctions as a longer-term hedging market.
It may also indicate limited confidence in projecting exchange rates and funding needs over longer periods.
A forward market is most useful when economic agents can plan ahead and hedge future obligations. If nearly all demand clusters at the short end, the auction may be functioning more as delayed spot intervention than as a deep market for managing longer-term currency risk.
The paper calls for further research into why forex users showed so little interest in longer maturities.
The regular forward-auction programme became increasingly difficult to maintain in 2022 after Ghana lost access to international capital markets.
Rising external debt obligations, portfolio-investor rollover risks and persistent demand for trade financing placed pressure on the Bank of Ghana’s reserves.
The central bank suspended the broader forward auctions in October 2022 and prioritised foreign exchange for the importation of petroleum products through auctions restricted to licensed bulk oil distributors.
The shift reflected a fundamental limit on foreign-exchange intervention: a central bank can only sell dollars it possesses or can reliably obtain.
When reserves become scarce, the question changes from how to allocate foreign exchange efficiently to which sectors should receive it first.
Oil imports were prioritised because fuel shortages or rapid increases in petroleum prices could spread through transportation, food distribution and broader inflation.
But sector-specific allocation also carries risks. Preferential access can fragment the foreign-exchange market, obscure its true clearing price and transfer currency risk from selected industries to the central bank.
Ghana’s first forex auction was introduced in September 1986 to narrow an extraordinary gap between the official and parallel-market exchange rates.
That system began as a retail auction, evolved into a wholesale arrangement in March 1990 and was discontinued in March 1992 as foreign-exchange allocation moved to the interbank market.
Its history shows that auction design has continually changed in response to Ghana’s economic conditions.
The first episode sought to replace administrative allocation and multiple exchange rates. The second sought to intermediate cocoa proceeds transparently and support price formation. The third attempted to reduce demand front-loading and provide certainty about future foreign-exchange supply.
Each addressed a different market weakness. None could remove the deeper constraint created when demand for dollars persistently exceeded available inflows.
The study recommends that any future auction targets should better reflect prevailing market demand and that auction guidelines should avoid price limits that obstruct genuine price discovery.
But implementing that recommendation would require policymakers to accept a trade-off.
Offering enough foreign exchange to satisfy demand can reduce immediate pressure on the cedi, but may rapidly deplete reserves if the underlying demand is structural. Allowing the price to adjust freely can preserve reserves and improve market clearing, but may produce a politically and economically painful depreciation.
The solution cannot therefore rest on auction design alone.
A sustainable foreign-exchange market requires stronger export receipts, predictable portfolio and direct-investment inflows, disciplined fiscal and monetary policies, and credible management of external debt obligations.
The Bank of Ghana paper is formally research in progress, and its conclusions do not necessarily represent the views of the central bank’s Board or management.
Its evidence nevertheless provides a timely warning: auctions are allocation mechanisms, not substitutes for foreign-exchange supply.
When the market requires three dollars for every one the central bank can provide, the problem is not simply how the auction is conducted. It is the underlying scarcity the auction exposes.
