The year 2020 was a particularly difficult one for economic and political forecasting. In a retrospective report, “Predicting the unpredictable: Forecasting in 2020”, we asked our analysts how they approached forecasting in a year like no other.

In the report, senior analyst Abhijit Surya explains how we refined our modelling skills and developed our use of alternative data to beat consensus with our Latin America forecasts.

How did Latin America’s economy fare in 2020?

We were pessimistic about Latin America’s prospects early on, and rightly so; the region was among the worst hit in the world from both an economic and a public health standpoint. However, in many countries the resilience of the external sector surprised us. This reflected high levels of support from multilateral institutions, strong Chinese demand for commodities exports and the fact that remittances inflows also held up well against all odds.

How did we adapt our approach to forecasting Latin America throughout the year?

As the coronavirus crisis escalated, our modelling skills came to the fore. We did a few important things that helped us get our forecasts right for Latin America. The first, building on our strengths developing indexes, was to create a heatmap assessing the vulnerability of the region’s economies to Covid-19, based on health indicators, economic structure indicators (like the level of labour informality) and scope for policy stimulus. The second step that we took was to complement our demand-based GDP forecasting model with a detailed analysis of GDP by sector, which allowed us to assess industry-by industry the impact of lockdowns and demand collapse in the region. And finally, we deepened our use of alternative, real-time data to get a stronger handle on the immediate effect of lockdown restrictions. Apart from mobility statistics, we followed daily electricity demand data closely, and this proved very useful.

What was the result of this change in approach?

All this did not prevent us from having to revise down our forecasts a couple of times, but we did adjust our projections faster than other forecasters, and our Q2 and Q3 GDP forecasts were spot on for most of the big economies in Latin America. Looking ahead, we have built another factor into our forecasts for 2021 and beyond: vaccine rollout.

For an in-depth look at the trials and tribulations of forecasting throughout 2020 in both Latin America and the rest of the world, download our latest report, “Predicting the unpredictable: Forecasting in 2020”.

All our award-winning forecasts can be found in our Country Analysis service. With detailed reports available for more than 190 countries around the world, our insights provide truly global coverage to help you to prepare for what’s ahead.

Abhijit Surya is a senior analyst in the Latin America team, responsible for economic and political analysis of various countries in the region. He contributes to The Economist Intelligence Unit’s extensive product portfolio on Latin America, which includes country reports, macroeconomic forecasts, and sovereign credit and operational risk assessments. His research interests include public finance, labour policy and the study of economic and financial crises.