- Bank Of Ghana Turns To AI, Big Data and Machine Learning to Sharpen Economic Forecasting
The Bank of Ghana is increasingly deploying artificial intelligence, Big Data and machine-learning technologies to generate faster intelligence on inflation, economic activity and financial-sector risks, signalling a significant shift towards real-time, technology-driven policymaking at the central bank.
First Deputy Governor Dr Zakari Mumuni said the Bank has developed an in-house electronic inflation nowcasting methodology known as e-Inflation, while machine-learning models are being used alongside conventional econometric models to forecast gross domestic product and undertake text-mining analysis.
The technologies are intended to narrow the gap between developments in the economy and the point at which policymakers become aware of them, allowing the Bank to identify emerging pressures earlier and respond more effectively.
Dr Mumuni disclosed the initiatives in a keynote address delivered at the 4th Annual Statistics and Data Science Conference 2026 of the Ghana Statistical Association in Tamale on Wednesday, August 26. The conference was held under the theme, “Innovations in Statistics and Data Science: Research, Practice and Policy Impact.”
“The greatest challenge facing policymakers today is no longer a shortage of data. It is turning an abundance of data into timely, reliable and actionable intelligence,” he said, arguing that the usefulness of data ultimately rests on its ability to improve decisions and outcomes.
The Bank’s growing use of AI represents a potentially important transformation in how monetary policy is formulated in Ghana. Traditional macroeconomic indicators are often released with a lag, meaning policymakers can sometimes be responding to conditions that have already shifted.
By combining conventional statistical systems with higher-frequency information and new analytical tools, the central bank is seeking to obtain a more immediate reading of economic conditions.
Dr Mumuni said the Bank’s inflation-targeting framework remains forward-looking and requires policymakers to assess where the economy currently stands, where it is heading, what could alter that trajectory and the appropriate policy response.
That process draws on information covering prices, output, credit, exchange rates, fiscal conditions and financial markets, with the Bank using econometric techniques and its Quarterly Projection Model within a Forecast and Policy Analysis System to assess risks and possible policy outcomes.
Technology is increasingly being layered onto that framework.
“The Bank has deployed AI and Big Data technologies to develop its in-house electronic inflation nowcasting methodology — e-Inflation,” Dr Mumuni said.
It is also using machine-learning techniques to complement standard models used in forecasting GDP, alongside text-mining analytics capable of extracting useful signals from large volumes of information.
The changes are extending into financial-sector supervision.
According to the First Deputy Governor, supervision previously relied heavily on static monthly spreadsheets that required manual reconciliation. Increasingly granular data can now be validated as it arrives, enabling supervisors to identify vulnerabilities earlier rather than discovering problems after they have become significantly larger.
That evolution could become increasingly important as Ghana’s financial system becomes more digital and transactions generate greater volumes of usable information.
Dr Mumuni raised the possibility of drawing intelligence from digital payment activity to obtain faster indications of consumption trends, using tax information as an early signal of business activity and deploying satellite imagery to assess agricultural conditions before conventional harvest statistics become available.
For monetary policymakers, such information could provide earlier warning of inflationary pressures, changes in consumer demand and supply-side disruptions.
But the Bank is also warning against treating technological sophistication as a substitute for the fundamentals of statistics.
Dr Mumuni stressed that AI cannot compensate for weak underlying data, arguing that unclear definitions, inconsistent classifications and inadequate validation can cause even sophisticated models to produce misleading results.
“Bad data in, very sophisticated garbage out,” he said, emphasising the continuing importance of sampling, measurement, classification, validation, metadata and revision.
The warning is significant as public institutions increasingly explore AI-based systems. The reliability of an algorithmic forecast ultimately depends on the quality, representativeness and governance of the information entering the model.
Dr Mumuni also argued that Ghana’s statistical systems must evolve as the economy changes. New industries emerge, consumption patterns shift and classifications or price baskets designed around the structure of the economy a decade earlier can become increasingly inaccurate if they are not revised.
He cited the rebasing of GDP and the Consumer Price Index as examples of why statistical frameworks must remain relevant, while noting that new data sources from payments, telecommunications and taxation create opportunities but also raise confidentiality and governance responsibilities.
The Bank’s technology push does not, however, represent an attempt to automate monetary policymaking.
Dr Mumuni said AI must strengthen rather than replace human judgment, reiterating a commitment made by Governor Dr Johnson Pandit Asiama when he took office in February 2025 to pursue a more proactive and precise approach to inflation management using advanced analytics and artificial intelligence.
“Technology can strengthen our intelligence, but it does not remove the need for human judgment,” the First Deputy Governor said. “People make policy.”
He also called for closer collaboration between policymakers and academia, arguing that Ghana’s universities possess substantial research capacity while public institutions simultaneously face urgent practical questions.
Research, he said, should increasingly be undertaken “with policy” rather than merely “about policy”, requiring researchers to understand the problems confronting public institutions and policymakers to remain open to evidence that challenges established assumptions.
Dr Mumuni identified three priorities for Ghana’s emerging data ecosystem: improving data quality, embracing high-frequency and alternative data without abandoning established statistical standards, and strengthening the connection between research and practical policymaking.
The Bank of Ghana’s development of e-Inflation and growing deployment of machine learning suggest that this transformation is already moving from theory into monetary-policy practice.
For the central bank, the objective is not simply to possess more data or more powerful algorithms. It is to identify economic changes sooner, understand them more accurately and give policymakers better evidence before decisions are made.
As Dr Mumuni put it, the ambition is ultimately straightforward: “Better data. Better decisions. Better outcomes.”
