Home / Current Issue / Paper 1713628
Real-Time Macroeconomic Now Casting Using High-Frequency and Non-Traditional Data
Subject area: Science,Engineering and Technology · Area of research: Macroeconomic Now Casting
DOI: https://doi.org/10.64388/IREV6I11-1713628
Abstract
Real time evaluation of macroeconomic conditions is essential in decision making in both the policy making and financial decision making but traditional indicators tend to be lagged and low-frequency, which limits their responsiveness. In this work, the authors use high-frequency financial and economic data, as well as non-traditional data, including the search trends in internet search engines and social media sentiment, to create a sophisticated nowcasting model of near-term variations of the macroeconomic variables. The approach combines a variety of data streams in real time to improve the predictive accuracy using a mixture of both econometric, as well as machine learning models. Empirical evidence shows that the use of high-frequency data and alternative data has a significant effect in improving performance on nowcasting relative to classic techniques, which minimises inaccuracies in prediction and improves the timeliness of economic information. The results as the policy makers, central banks, and financial analysts can use to transform conventional economic signals to alternative data sources that can support economic metrics. The paper has shown that real time nowcasting of macroeconomic variables can be more accurate and responsive and therefore make decisions in volatile economic situations faster and better informed.
Keywords
Macroeconomic Nowcasting, High-Frequency Data, Non-Traditional Data, Real-Time Forecasting, Machine Learning
How to cite this paper
@article{1713628,
author = {Oksana Anatolyevna Malysheva},
title = {Real-Time Macroeconomic Now Casting Using High-Frequency and Non-Traditional Data},
journal = {Iconic Research And Engineering Journals},
year = {2023},
volume = {6},
number = {11},
pages = {1020-1030},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1713628.pdf},
abstract = {Real time evaluation of macroeconomic conditions is essential in decision making in both the policy making and financial decision making but traditional indicators tend to be lagged and low-frequency, which limits their responsiveness. In this work, the authors use high-frequency financial and economic data, as well as non-traditional data, including the search trends in internet search engines and social media sentiment, to create a sophisticated nowcasting model of near-term variations of the macroeconomic variables. The approach combines a variety of data streams in real time to improve the predictive accuracy using a mixture of both econometric, as well as machine learning models. Empirical evidence shows that the use of high-frequency data and alternative data has a significant effect in improving performance on nowcasting relative to classic techniques, which minimises inaccuracies in prediction and improves the timeliness of economic information. The results as the policy makers, central banks, and financial analysts can use to transform conventional economic signals to alternative data sources that can support economic metrics. The paper has shown that real time nowcasting of macroeconomic variables can be more accurate and responsive and therefore make decisions in volatile economic situations faster and better informed.},
keywords = {Macroeconomic Nowcasting, High-Frequency Data, Non-Traditional Data, Real-Time Forecasting, Machine Learning},
month = {May},
doi = {https://doi.org/10.64388/IREV6I11-1713628}
}