Economists are increasingly turning to new technologies that can track indicators such as growth and inflation in real time to refine their forecasts and provide better insights to policymakers.
Nowcasting, or forecasting of the present, is particularly promising for developing economies where statistical authorities do not often publish indicators. At the IMF, we have developed an approach that combines high-frequency data with machine learning, a kind of artificial intelligence, to provide nowcasts of economic growth and help policymakers make better decisions.
Some developing economies publish key indicators such as gross domestic product with long delays. This adds to the challenges of policymaking in times of rapid change, such as the onset of the pandemic or the outbreak of war in Ukraine, as decisions have to be made without crucial data. At times like these, the nowcast can offer projections that approximate economic activity much faster than official GDP data.
As the Chart of the week shows, the nowcasting framework offered invaluable insights for Botswana when the pandemic shutdowns hit in the second quarter of 2020. At the time, it forecast production to contract by around 20% year-on-year. former. In September, when the government released GDP figures, they showed even sharper growth 24% contraction. It wasn’t far off from the frame’s projections, which tracked the actual data and accurately predicted the turn.
In a new journalwe seek to bridge the gap between data availability and policy-making in sub-Saharan Africa by developing a framework to track economic activity in real time.
Our nowcasting framework extracts signals from the high-frequency numbers available ahead of the official GDP release.
The tool generates nowcasts by incorporating a wide range of increasingly popular machine learning techniques that use an array of high-frequency economic indicators historically linked to the evolution of GDP. Tourist arrivals, for example, are more reliable predictors for tourism-dependent countries. For oil-exporting countries, like Nigeria, GDP tends to move with crude oil prices.
Other non-traditional data inputs may include satellite images of night lightswhich tend to shine with greater intensity as economic activity increases, and transport ships, used to track trading volumes and disruptions. Web searches can also help predict tourist arrivals more accurately, as one IMF working paper shown in 2020.
Our research shows that machine learning algorithms are often more accurate than traditional econometric methods, especially at predicting turning points such as when an economy begins to rebound from a recession or crisis.
The nowcasting predictors we used for our Botswana work included currency values, stock prices, inflation, imports, consumer loans, power generation, tax revenue, and volume. Google search for the country name (an indicator of future tourist visits). And because it is one of the largest exporters of diamonds in the world, our framework also incorporates stone prices as well as the economic growth of two of the main destination countries for their shipments.
The Bank of Botswana currently produces own nowcasts, after participating in an IMF workshop. Related courses on the development of high-frequency indicators will be offered to help authorities around the world better track economic activity.
*About the authors:
- Seung Mo Choi is a senior economist working on regional surveillance in the African Department of the IMF. He has worked on banking crises, financial market policies, climate change, low-income country issues, and capacity development, including at the European Department of the IMF and the Institute for Capacity Development. His research has been published in economic and financial journals such as International Economic Review.
- Tara Iyer is an Economist in the Global Financial Stability Analysis Division of the IMF’s Monetary and Financial Markets Department, where she contributes to the Global Financial Stability Report. Tara’s current research focuses on issues related to mutual funds, crypto assets, and monetary policy in emerging markets.
Source: This article was published by IMF Blog
