Nowcasting Swedish GDP with a large and unbalanced data set

Abstract

We evaluate pseudo-real-time out-of-sample nowcasts for Swedish GDP employing factor models and mixed-data sampling regressions with single predictor variables. These two model classes can handle the data irregularities of a ragged-edge sample and differing sampling frequencies. The results show that pooling of the nowcasts outperforms a simple benchmark, even though only very few of the underlying specifications achieve improved accuracy individually. Moreover, we assess the accuracy of the density forecasts, i.e., the uncertainty around the point forecasts. The post-crisis period after 2008 turns out to be a more difficult period to nowcast precisely. However, indicator variables prove more useful post-crisis as then the performance relative to univariate benchmarks improves.

Forecasting GDP with many predictors in a small open economy: forecast or information pooling?

“…Applications to small open economies are found in more recent studies such as Rusnák ( 2016 ) for Czech; Kim and Swanson ( 2018 ) for Korea; Yau and Hueng ( 2019 ) for Taiwan; Galli et al. ( 2019 ) for Switzerland; den Reijer and Johansson ( 2019 ) for Sweden; Laine and Lindblad ( 2021 ) for Finland; and Marcellino and Sivec ( 2021 ) for Luxembourg.…”

Bootstrap aggregation accuracy gains in nearest‐neighbor nowcasting of Swedish gross domestic product

“…This so-called Economic Tendency Survey data are often used when nowcasting both macroeconomic quantities in general 1 and specific quantities such as Swedish GDP growth. [2][3][4][5][6] It has turned out that machine learning algorithms can be useful for nowcasting macroeconomic aggregates. 7,8 When individual learning algorithms are unstable, that is, to a large degree affected by the specific data sample at hand, prediction results can be improved by averaging over individual learners based on bootstrap samples from the original data set.…”