What does this research mean for the field?

Machine learning models consistently outperform benchmark models in out-of-sample forecasts of regional real GDP growth, particularly in smaller regions with high economic volatility, provided dimension reduction is used to handle noisy big data. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

This research aims to evaluate the effectiveness of machine learning models in forecasting GDP growth using big data.

June 7, 2026

Regional Economic Forecasting with Big Hierarchical Data and Machine Learning Models

Key Points

This research aims to evaluate the effectiveness of machine learning models in forecasting GDP growth using big data.
Utilized provincial data from Guangdong and municipal data from Shenzhen, China.
Compared machine learning models with benchmark models for out-of-sample GDP growth forecasts.
Employed dimension reduction techniques to manage additional noisy national-level data.
Machine learning models outperformed benchmark models in forecasting real GDP growth, especially in smaller, volatile regions.
Critical predictive variables for economic conditions were identified, showcasing the importance of model-driven approaches.
Periodical reexamination of informative data sets is necessary as the relevant data may change over time.

Abstract

Abstract Big data offers policy makers a wealth of real-time information, but distinguishing valuable information from noise remains a significant challenge. Using a large set of provincial data from Guangdong, China, and municipal data from Shenzhen, we find that machine learning models consistently outperform benchmark models in out-of-sample forecasts of real GDP growth, particularly in smaller regions with higher economic volatility. Next, we test performance of the models with additional national-level data and found that dimension reduction is critical for the models to utilize additional noisy data. Finally, a model-driven approach can effectively identify critical predictive variables that policy makers can use to monitor economic conditions. However, the set of informative data may shift over time, necessitating periodical reexamination using big data.

Ask AI

Mark Helpful

Bookmark

Relay