IMF Research Reveals How Advanced Economic Models Could Reshape Interest Rate Policy Worldwide

An IMF study introduces a modern Bayesian forecasting framework that improves central banks' ability to assess inflation, interest rates and economic growth while reducing uncertainty in monetary policy decisions. Tested on Brazil's economy, the model delivers more stable forecasts and offers policymakers, businesses and financial institutions better tools to manage economic shocks and financial risks.

The International Monetary Fund (IMF) has introduced a new economic forecasting framework that could help central banks make better decisions on inflation, interest rates and economic growth. In a 2026 working paper titled From Calibration to Estimation: A Modern Bayesian Framework for Quarterly Projection Models, researchers Andrés González Gómez and Juan David Prada from the IMF Research Department explain how advanced statistical methods can improve monetary policy analysis. Using Brazil's economic data from 2004 to mid-2025, the study demonstrates how central banks can reduce their dependence on predetermined assumptions and develop more reliable economic forecasts. The findings have important implications for governments, businesses, financial institutions and consumers facing uncertain economic conditions.

A New Approach to Predicting Economic Changes

Central banks use Quarterly Projection Models to understand how inflation, interest rates, exchange rates and economic growth influence one another. These models help policymakers predict economic developments and assess the possible consequences of monetary policy decisions.

However, traditional models often rely on assumptions selected by economists rather than parameters fully estimated from economic data. While this approach makes forecasting easier, it can overlook changing economic conditions and create uncertainty about the actual effects of policy decisions.

The IMF researchers propose a modern Bayesian framework that combines economic theory with historical data to improve forecasting accuracy. Their approach uses advanced computational techniques, including the structure-preserving doubling algorithm, automatic differentiation and the No-U-Turn Sampler.

These methods allow economists to estimate numerous economic relationships simultaneously while measuring uncertainty more effectively. The framework also separates temporary economic shocks from longer-term trends, helping policymakers distinguish short-lived disturbances from structural changes.

The researchers tested ten model specifications using 15 economic indicators from Brazil, including GDP, consumer inflation, food and energy prices, interest rates, exchange rates and sovereign risk premiums. Their preferred model estimates 71 parameters and tracks 39 economic states.

Brazil's Inflation Battle Reveals Strong Monetary Policy Response

One of the study's most important findings concerns how Brazil's central bank responds to inflation. The estimated inflation-response coefficient reaches approximately 3.38, indicating a strong systematic response to expected inflationary pressures.

The model also identifies substantial interest-rate persistence, with a policy-smoothing coefficient of approximately 0.92. This suggests that monetary authorities tend to adjust interest rates gradually while maintaining a firm commitment to controlling inflation.

The research highlights the importance of global commodity markets in shaping Brazil's economic performance. Commodity prices measured in US dollars show a correlation of 0.71 with Brazil's output gap, indicating a strong relationship between international prices and domestic economic activity.

However, this relationship weakens considerably when commodity prices are measured in Brazilian reais. The finding suggests that exchange-rate flexibility can absorb part of the impact of international commodity-price shocks.

For traders and businesses, these relationships are particularly important. Currency depreciation can increase import costs, raise production expenses and affect profit margins. Consumers may experience these pressures through higher food prices, energy bills and other essential household expenses.

Interest Rates and Inflation Control Come at an Economic Cost

The study provides important estimates of Brazil's neutral real interest rate, which represents the inflation-adjusted interest rate consistent with sustainable economic activity without generating additional inflationary pressure.

Researchers estimate that this rate declined from approximately 10% during 2004–2005 to around 5% by 2014, reaching 5.11% by mid-2025.

Understanding this benchmark can help policymakers determine whether monetary policy is restricting economic activity or supporting growth. Interest rates above the neutral level generally discourage borrowing and investment, while lower rates can stimulate economic demand.

The research also examines the economic cost of reducing inflation. Its estimated sacrifice ratio reaches 0.37, highlighting the trade-off between price stability and economic activity under the model's assumptions.

Another significant finding concerns the output gap, which measures the difference between actual economic production and the economy's estimated potential. The model's output-gap estimates fall within the range published by the Banco Central do Brasil in 83.7% of the quarters examined.

The model also captures major economic disruptions, including the global financial crisis, Brazil's 2014–2016 recession and the COVID-19 pandemic. During the second quarter of 2020, the estimated output gap fell to approximately minus 7.6 percentage points.

Better Forecasts Could Strengthen Economic Policymaking

The IMF framework substantially reduces uncertainty surrounding monetary policy transmission. At a four-quarter horizon, uncertainty ranges around estimated policy responses become five to twelve times narrower than those generated by initial calibrated assumptions.

Forecasting comparisons also show that the estimated model produces relatively stable predictions for inflation, economic growth and interest rates, with its clearest advantage involving Brazil's Selic policy rate.

However, researchers acknowledge that the forecasting comparison is not a fully realistic real-time competition because the projection model uses parameters estimated from the complete dataset.

The study also identifies challenges in modelling extraordinary disruptions such as the COVID-19 pandemic. Future improvements could incorporate fiscal policy, changing economic volatility and more detailed commodity-market relationships.

For governments and development institutions, the findings suggest opportunities to strengthen economic forecasting and improve policy planning. Financial institutions and businesses could benefit from clearer assessments of interest-rate movements, inflation risks and economic conditions.

The broader message is that combining economic theory with modern statistical estimation can help central banks make more informed decisions. Although the framework cannot eliminate economic uncertainty, it offers a practical way to improve monetary policy analysis, support price stability and reduce the risks associated with poorly informed economic decisions.