Bank of Ghana Incurred $1.7 Billion Losses on Gold Programme Involving GoldBod in 2025: IMF Report
The International Monetary Fund (IMF) has disclosed that the Bank of Ghana (BoG) recorded losses of more than US$1.7 billion in 2025 through its Domestic Gold Purchase Programme (DGPP), particularly under the Gold-for-Reserves operations involving the Ghana Gold Board (GoldBod).
According to the IMF’s Selected Issues Report released in early August 2026, the significant scaling up of the programme led to losses equivalent to 1.5 percent of GDP. The losses were almost entirely related to purchases of doré gold from artisanal and small-scale miners.
The report notes that the BoG purchased and exported around 104 tonnes of artisanal gold worth approximately US$10.9 billion in 2025, making the central bank the main channel for nearly all officially recorded artisanal gold exports.
While the programme contributed to macroeconomic stabilisation helping boost foreign exchange reserves to US$11.9 billion and supporting a strong appreciation of the cedi, it came at a high cost. Losses stemmed from a combination of service and assay fees paid to GoldBod, discounts on gold sold to off-takers, and exchange rate differences between the rates paid to purchase gold and the BoG’s accounting reference rate.
Losses in 2024 had been about US$400 million. The sharp rise in 2025 reflected the expanded scale of operations.
The IMF observed that these losses amounted to roughly 17 percent of the value of the doré gold sold by the BoG.
In July 2026, responsibility for the programme was fully transferred to GoldBod under a memorandum of understanding. The Bank of Ghana has exited the arrangement and is no longer exposed to the associated quasi-fiscal financial risks. Under the new structure, transaction costs are expected to decline significantly.
The findings form part of the IMF’s broader assessment of Ghana’s gold sector policies and their impact on the central bank’s balance sheet and overall macroeconomic management.
Editor:
Obiri-Yeboah
