A1 Vertaisarvioitu alkuperäisartikkeli tieteellisessä lehdessä
Forecasting With Dynamic Factor Models Estimated by Partial Least Squares
Tekijät: Rauhala, Samuel
Kustantaja: Wiley
Julkaisuvuosi: 2026
Lehti: Journal of Forecasting
Artikkelin numero: for.70158
ISSN: 1099-131X
eISSN: 0277-6693
DOI: https://doi.org/10.1002/for.70158
Julkaisun avoimuus kirjaamishetkellä: Avoimesti saatavilla
Julkaisukanavan avoimuus : Osittain avoin julkaisukanava
Verkko-osoite: https://doi.org/10.1002/for.70158
Rinnakkaistallenteen osoite: https://research.utu.fi/converis/portal/detail/Publication/523225592
Rinnakkaistallenteen lisenssi: CC BY
Rinnakkaistallennetun julkaisun versio: Kustantajan versio
Dynamic factor models (DFMs) have found great success in nowcasting and short-term macroeconomic forecasting when incorporating large sets of predictive information. The factor loadings are typically estimated cross-sectionally with principal component analysis (PCA) or maximum likelihood (ML), which ignore whether the factors have predictive power. We suggest two novel alternative approaches using partial least squares to estimate large vector autoregressions (VARs) and DFMs, which take the dynamic dependencies better into account. Our Monte Carlo simulations and forecasting results for the Finnish GDP growth show that these methods generally perform on par with and under certain conditions better than the existing approaches.
Ladattava julkaisu This is an electronic reprint of the original article. |  |
Julkaisussa olevat rahoitustiedot:
The author would like to thank Petteri Juvonen, Juho Koistinen, Markku Lanne, Henri Nyberg, Joni Virta, and seminar participants at the Bank of Finland (2024), Nordic Econometric Meeting (Bergen 2024), CFE 2024 (London), and Helsinki Graduate School of Economics (2024). The financial support from the OP Group Research Foundation (grant 20230116), the Foundation for Economic Education (Liikesivistysrahasto, grant 220246), the Turku University Foundation (Turun Yliopistosaatio, grant 081875), and The Finnish Doctoral Program Network in Artificial Intelligence, AI-DOC (decision number VN/3137/2024-OKM-6) is gratefully acknowledged. Special thanks are due to the Bank of Finland for the data used in Section 5 and research cooperation (research visit in the autumn 2024). Open access publishing facilitated by Turun yliopisto, as part of the Wiley - FinELib agreement.
