By J. Atsu Amegashie
There have been vibrant debates about the loss of $214 million (reported by the IMF) associated with the domestic gold purchase program (DGPP). The bone of contention is whether the loss should be on the books of the Bank of Ghana or the government of Ghana.
The IMF wants the loss to be on the budget of the government of Ghana (represented by the GoldBod). The CEO of GoldBod, Sammy Gyamfi, disagrees. Some economists have argued that central bank losses should be viewed as the result of quasi-fiscal activities or a form of fiscal policy.
Based on trade reports published by GoldBod, the Bank of Ghana gave an amount of 112,654,492,415.60 cedis (about 112.65 billion cedis) to the GoldBod. After buying and exporting gold, the GoldBod, acting as an agent of the BoG, sent an amount of 109,145,172,275.50 cedis (about 109.14 billion cedis) to the Bank of Ghana, a net amount of about 3.5 billion cedis to the BoG.
I interpret “including undisbursed funds” as undisbursed funds from the GoldBod to the BoG. In that case, the net income of 3.5 billion cedis is a net loss to the BoG.
On JoyNews Newsfile programme on January 3, 2026, Sammy Gyamfi said that over 97% of the BoG’s reported loss under the domestic purchase program (which includes gold for reserves) was the result of the fact that the GoldBod bought gold at retail market exchange rates, but the proceeds of gold exports were valued at BoG reference exchange rates or interbank rates that were lower than the retail market rate.
He said also that, “… an accountant can say you have made a loss. But it is an intentional, strategic, economic policy cost for superior benefits … Under the gold for reserves (G4R) program, cost is the necessary evil for national economic stability. … The BoG is not a gold trading organization. … Under the new GoldBod model … GoldBod will be the buyer and seller because currently we are not the seller. BoG is the seller. We are just their agent and we are paid fees for that. … Any reported loss under G4R … is a product of policy design. That was how the policy was designed to incur intentional and strategic expenses for superior economic benefits.”
The superior economic benefits, according to Sammy Gyamfi, were “… $10.8 billion in foreign exchange. 40% appreciation of the cedi. We have brought the retail dollar rate from about 16 cedis to below 12 cedis. Our external debt has been reduced by over 100 billion cedis. We saved 16 billion cedis in debt servicing … Inflation has declined for 11 consecutive months from 23.8% to 6%. We have met our foreign reserves (under the IMF program) three years ahead of time.”
Of course, one cannot attribute all these achievements to only the GoldBod. Still, they are impressive achievements. But should we engage in a cost-benefit analysis where “any reported loss … is a product of policy design … to incur intentional and strategic expenses for superior economic benefits”?
Imagine that the GoldBod was not an agent of the BoG. Suppose it paid 12 billion cedis to artisanal and small-scale miners to buy $1 billion worth of gold at $1 = 12 cedis. Suppose the GoldBold exported the gold and got $1.1 billion. It handed this amount to the Bank of Ghana at a rate of $1 = 10 cedis. The GoldBod got 11 billion cedis, so it made a loss of 1 billion cedis.
In this example, the GoldBod bought dollars (gold priced in dollars) at a relatively expensive price (i.e., 12 cedis), but sold dollars to the BoG at a cheaper price (i.e., 10 cedis). It is a loss caused by an appreciation of the cedi from $1 = 12 cedis to $1 = 10 cedis during the period between the purchase of gold and the sale of gold. The appreciation of the cedi may be natural (i.e., if the rate was determined by market forces) or artificial (e.g., if it was artificially set by the BoG). Still, the GoldBod’s loss is real: an outflow of 12 billion cedis and an inflow of 11 billion cedis.
In contrast, suppose this transaction was undertaken by the Bank of Ghana, and the loss was put on the books of the Bank of Ghana. The BoG bought $1 billion worth of gold and sold it for $1.1 billion, a surplus of $0.1 billion = $100 million. The $1.1 billion will be in the “vaults” of the Bank of Ghana. In cedis, the Bank of Ghana, may have made a loss. But it may not matter because the BoG prints cedis.
Inflation is the real constraint on the BoG’s ability to print cedis. If it can print cedis without triggering a significant rise in the inflation rate, then it is in the best of both worlds. In contrast, a loss of 1 billion cedis on the books of the GoldBod requires the GoldBod to borrow money or significantly increase its future profits to offset the loss. The GoldBod does not and cannot print cedis.
In the same vein, arguing that losses by the ECG should be weighed against benefits like the increase in GDP, employment, and education as a result of the ECG’s supply of electricity to homes, businesses, universities, and schools does not change the fact that ECG must increase revenue, reduce costs, etc, to offset the loss. Much of the increase in GDP, employment, and education does not accrue to the ECG.
When a government spends on roads, hospitals, schools, etc, this usually has significantly positive effects on the economy and also yields non-monetary benefits (e.g., good health, low levels of stress in traffic, etc).
However, if the government is unable to capture a sufficient percentage of the increase in GDP through tax revenue (i.e., a low tax buoyancy) and is unable to borrow at reasonable interest rates, it may be forced to default on the loans that financed the GDP-boosting physical infrastructure, resulting in disastrous economic effects.
For the GoldBod (in 2026 and beyond), it has to be profitable when every aspect of its full operations is taken into account. Otherwise, it may not have the money to buy gold, and its creditors/suppliers could take legal action against it.
At a meeting in November 2020, Christine Lagarde, the President of the European Central Bank (ECB) and former Managing Director of the IMF, was asked the following question:
“… the ECB could incur losses related to its holdings under the asset purchase programme (APP). So I would like to know, technically, what would happen if those losses were to erode the equity of the ECB and how it is possible that the ECB could also run with negative equity.?”
Christine Lagarde’s response was:
“As the sole issuer of euro-denominated central bank money, the euro system will always be able to generate additional liquidity as needed. So by definition, it (i.e., the ECB) will neither go bankrupt nor run out of money. … We are the only issuer (of the euro), and we are not at risk as a result.” Parentheses mine.
Thus, in the case of the BoG, we have the luxury of engaging in broad macro cost-benefit analysis because, as the sole issuer of cedis, the capacity of the BoG to absorb cedi-denominated losses is much bigger than the capacity of the GoldBod, ECG, etc. The BoG does not have to print money to cover its loss. However, the point is that the BoG has the “money-printing” tool in its arsenal of options. From the standpoint of the BoG, we can justify the loss in cedis on grounds that inflation decreased, the cedi appreciated in value, and the cedi-dollar exchange rate was stabilized, the cost of external debt service fell, etc.
The BoG can also lose money in dollars, a currency that it cannot print. For example, this may occur if the BoG enters into a contract to deliver dollars (e.g., to commercial banks) in the future (a forward contract). Based on Sammy Gyamfi’s claim that over 97% of the BoG’s losses were the result of exchange rate differentials, I ignore dollar-denominated losses.
The BoG can continue its domestic gold purchase programme or G4R programme in 2026, although the GoldBod has fully taken over the purchase and sale/export of gold. The central banks of countries like Germany, France, Italy, Japan, the UK, India, Turkey, Poland, Singapore, etc have gold reserves, but these countries are not gold producers. They buy gold from the world market. When the price of gold falls, its value (in dollars) in the BoG’s reserves will fall. That is a loss. But that loss, related to reserve management and monetary policy, is different from the loss of actually trading in gold.
Sammy Gyamfi has admitted that the BoG is not a gold trading organization. Gold trading is not a part of the BoG’s mandate. Central bank losses usually originate when the bank takes on other functions _besides its normal role_, such as subsidized loans to priority sectors (housing, agriculture, strategic industries, exports), rescues of troubled financial institutions, or takeovers of private or public debt.
There are well-known justifications for the independence of central banks and restrictions on their capacity to directly lend or give monetary transfers to fiscal authorities (the government). An old, non-mainstream, and resurrected theory in economics — renamed as modern monetary theory (MMT) — challenges this view.
The essential message of MMT is that there is no financial constraint on government spending as long as a country is a sovereign issuer of currency and does not tie the value of its currency to another currency. Thus, central banks of countries that issue their own (fiat) currency can print as much money as the government needs, particularly when unemployment is high.
Thus, from the standpoint of MMT, the main fiscal constraint is not where to get money but whether those newly created units of money can be absorbed by the economy’s spare (unused) capacity without creating unsustainable inflation. MMTers argue that printing money to finance government expenditure is desirable IF it is directed toward productivity-enhancing projects.
The major objection to MMT is that it will facilitate unlimited government spending that will ultimately result in an accelerating rate of inflation with damaging consequences for economic growth. As an example, before Greece joined the eurozone, the Bank of Greece issued the domestic currency, the drachma. The Bank of Greece was at the beck and call of the Greek government, financing the government’s fiscal deficits. At one point during this period, the same person held the position of finance minister and governor of the Bank of Greece. The annual rate of money growth averaged more than 20% from 1981 to 1994, while the rate of inflation averaged 18%, a very high rate by OECD standards. Scholars have analyzed some of Latin America’s episodes with MMT-type policies in Chile, Peru, Argentina, and Venezuela.
In each country, populist governments financed fiscal deficits through money printing by the central bank. The consequence was hyperinflation and significant declines in consumers’ purchasing power. A study of 107 countries from 1960 to 2001 found a strong positive association between deficits and inflation among high-inflation and developing countries, but not among low-inflation advanced economies.
Given that the gold from artisanal and small-scale miners was paid for in cedis that were given to the GoldBod by the BoG, the relationship between the BoG and the GoldBod (a government agency) and the ensuing debates about the BoG’s loss reminded me of MMT. The BoG-GoldBod arrangement was, in effect, direct financing of government operations by the BoG, but with GoldBod labelled as an agent of the BoG.
However, the arrangement was not a classic MMT-type financing of general government expenditure. It was limited to gold purchases, a commodity that has a direct effect on the cedi-dollar rate via dollar inflows from exports and a disinflationary effect through lower cedi-equivalent import prices (i.e., exchange rate pass-through).
Given Ghana’s import-dependent economy, the favorable exchange rate pass-through was one of the reasons for the significant reduction in inflation in 2025. However, considering its history of central bank financing of government budget deficits and the high levels of inflation in the 1970s and early 1980s, Ghana has placed a significant restriction on the BoG’s monetary transfers to the government. We should be guided by history.
Thus, from the standpoint of fiscal and monetary discipline and the management of perverse incentives, the loss of $214 million in 2025 should be on the books of the GoldBod. It appears that the BoG will assume the loss. Going forward (i.e., 2026 and beyond), the GoldBod has taken full responsibility for all aspects of its operations. It is no longer an agent of the BoG. This is a better and incentive-compatible arrangement.
