Deep Impulse Response Functions for Macroeconomic Dynamics: A Hybrid LSTM-Wavelet Approach Compared to an ANN-Wavelet and VECM Models

Bahaa Aly, Tarek (2025): Deep Impulse Response Functions for Macroeconomic Dynamics: A Hybrid LSTM-Wavelet Approach Compared to an ANN-Wavelet and VECM Models.

Abstract

This study presents a novel hybrid framework that integrated Long Short-Term Memory (LSTM) networks with Daubechies wavelet transforms to estimate Deep Impulse Response Functions (DIRF) for monthly macroeconomic time series, across five economies: Brazil, Egypt, Indonesia, United States, and the United Kingdom. Eight key variables, yield curve latent factors (LEVEL, SLOPE, CURVATURE), foreign exchange rates, equity indices, central bank policy rates, GDP growth rates, and inflation rates, were modeled using the proposed LSTM-Wavelet approach, and were compared against an ANN-Wavelet hybrid, and a traditional Vector Error Correction Model (VECM). The LSTM-Wavelet model achieved a superior overall median R2, outperforming the ANN-Wavelet and VECM. The approach excelled in capturing nonlinear dynamics and temporal dependencies for variables such as equity indices, policy rates, GDP, and inflation. Db4 was superior for capturing short and medium-term patterns in macroeconomic variables like GDP, EQUITY, and FX, cause its shorter filter and moderate smoothing excelled at isolating cyclical patterns in noisy, volatile data. Cumulative DIRFs revealed consistent cross variable dynamics e.g., yield curve shocks propagated to equity, FX, policy rates, GDP, and inflation, in line with economic theory. These findings underscored the hybrid model’s ability to capture non-linearity, multiscale interactions in macroeconomic data, offering valuable insights for forecasting and policy analysis.

• Item Type: MPRA Paper
• Language: English
• Keywords: Deep Impulse Response Function, Long Short-Term Memory, Daubechies Wavelet transform, Macroeconomics, nonlinearity, Forecasting
• Subjects: C - Mathematical and Quantitative Methods > C5 - Econometric Modeling; C - Mathematical and Quantitative Methods > C5 - Econometric Modeling > C53 - Forecasting and Prediction Methods ; Simulation Methods; C - Mathematical and Quantitative Methods > C5 - Econometric Modeling > C58 - Financial Econometrics

References

Sezer, O. B., Gudelek, M. U., & Ozbayoglu, A. M. (2020). Financial Time Series Forecasting with Deep Learning: A Systematic Literature Review. Applied Soft Computing, 84, 105766. Chen, Z., Li, C., & Sun, W. (2020). Bitcoin Price Prediction Using Machine Learning: An Approach to Sample Dimension Engineering. Journal of Computational and Applied Mathematics, 365, 112395. Liu, J., Wang, J., & Zhang, Y. (2021). Hybrid Wavelet-LSTM Model for Economic Time Series Forecasting. Economic Modelling, 94, 123-135. Kim, H., & Lee, J. (2021). Wavelet-Based LSTM for Economic Time Series Analysis. Journal of Computational Economics, 39(4), 123-140. Barzegar, R., Aalami, M. T., & Adamowski, J. (2021). Short-Term Water Demand Forecasting Using Deep Learning and Wavelet Transform. Journal of Hydrology, 598, 126210. Parvini, N., Abdollahi, M., & Seifollahi, S. (2022). Two-Stage Forecasting Method Based on Discrete Wavelet Transform and LSTM Network. Finance Research Letters, 45, 102134. Chen, L., & Zhang, Y. (2023). Hybrid LSTM-Wavelet Models for Macroeconomic Forecasting. Quantitative Finance, 23(2), 245-260. Winarno, S., Usman, M., & Warsono, K. (2023). Application of Vector Error Correction Model and Impulse Response Function for Daily Stock Prices. IOP Conference Series: Materials Science and Engineering, 1289, 012056. Li, X., & Wang, J. (2024). Quarterly GDP Forecast Based on Coupled Economic and Energy Feature WA-LSTM Model. Frontiers in Energy Research, 12, 103456. Vivas, E., Allende-Cid, H., & Salas, R. (2025). Long Short-Term Memory Wavelet Neural Network for Renewable Energy Generation Forecasting. International Journal of Intelligent Systems, 40(1), 123-145.