https://doi.org/10.52903/wp2026374
A SURVEY OF MACROECONOMIC NOWCASTING APPLICATIONS
Stavros Degiannakis
Panteion University of Social and Political Sciences and Bank of Greece
Ioannis Krompas
Panteion University of Social and Political Sciences and National Bank of Greece
ABSTRACT
Nowcasting, broadly understood as the real-time assessment of current and near-term economic conditions, is a framework of growing importance that helps inform both economic policy decision for Governments and Central banks but also shape strategies of private institutions. In this paper, a dataset of macroeconomic nowcasting application papers is compiled and analyzed to identify the main research themes and gaps in the literature. Results find that literature is rather concentrated in nowcasting aggregate economic measures, specifically GDP which appears as the target variable in 74% of the papers examined. In terms of modelling frameworks, even though there is no consensus about the best-performing type of model, Dynamic Factor Models seem to be widely preferred as they were used in 37.5% of the nowcasting applications. However, Machine Learning applications are being increasingly used after the COVID-19 pandemic, coupled by an increased reliance in the use of alternative sources of data, which feature in an average of 51.7% of annual publications between 2020-2026, compared with 18.6% during 2010–2019. Despite advances in the field, important challenges remain, as illustrated not only by the difficulties many nowcasting models faced during periods of economic turbulence but also more recently, by the Atlanta Fed’s toolkit’s “GDPNow” nowcasting undershoot associated with differences in the classification of gold imports. Further challenges arise from the insufficiently explored limitations and potential biases of alternative data sources, as well as difficulties in acquiring historical data vintages necessary for conducting true real-time nowcasting evaluations. Against this background, future research should focus on: (i) continuing on extending nowcasting applications to underexplored variables and economies, (ii) examining the predictive properties and limitations of alternative data, and (iii) examine the potential usefulness of evaluation frameworks that combine real-time predictor vintages with final revised estimates of target variables to assess predictive performance against their underlying economic movements while avoiding look-ahead bias, so as to avoid misguiding policy makers during periods of substantial data revisions.
Keywords: Dynamic Factor, Forecasting, Macro econometrics, Mixed Frequency, Nowcasting
JEL-classifications: C53, C55, C82, E27, E37
Acknowledgments: The views expressed in this paper are those of the authors and do not necessarily reflect those of Bank of Greece and National Bank of Greece. The authors are responsible for any errors or omissions in this paper.
Correspondence:
Stavros Degiannakis
Economic Analysis and Research Department
Bank of Greece
El.Venizelos 21, 10250 Athens, Greece
Tel.: +30-2103202386
email: sdegiannakis@bankofgreece.gr