Forecasting inflation with Artificial Intelligence:
A comparative analysis
DOI:
https://doi.org/10.46661/rev.metodoscuant.econ.empresa.12447Keywords:
large language models, GPT, inflation forecasting, inflation expectations, natural language processingAbstract
Inflationary dynamics underscore the need for advanced methodologies to enhance forecasting accuracy. This paper explores the potential of Artificial Intelligence (AI) in generating short-term inflation forecasts for Argentina during the 2023-2024 period. The methodology leverages OpenAI’s GPT-4o Mini model, a Large Language Model (LLM), to produce conditional predictions by supplying historical Consumer Price Index (CPI) data and explicitly restricting its knowledge base to the forecast date. Additionally, forecasts are benchmarked against the inflation expectations survey conducted by Argentina's Central Bank, known as the Relevamiento de Expectativas de Mercado (REM). While predicting high inflation spikes remains challenging for both approaches, our results indicate that the AI model achieves comparable performance to REM for medium to low monthly inflation rates. For instance, for forecasts made at a given month t (e.g., August 2024) and evaluated across the subsequent seven forecast horizons when monthly inflation is around 4%, the Mean Squared Error (MSE) for GPT-4o Mini's median predictions was 0.90 and the Mean Absolute Error (MAE) was 0.85, closely aligning with REM's performance, which recorded an MSE of 0.68 and an MAE of 0.73.
Downloads
References
Ang, A., Bekaert, G., & Wei, M. (2007). Do macro variables, asset markets, or surveys forecast inflation better? Journal of Monetary Economics, 54(4), 1163-1212.https://doi.org/10.1016/j.jmoneco.2006.04.006
Banco Central de la República Argentina. (BCRA) (2024). Relevamiento de Expectativas de Mercado (REM). August 2024 Report. BCRA.
Bybee, L. (2023). Surveying Generative AI's Economic Expectations. arXiv Preprint. arXiv:2305. 02823.https://doi.org/10.2139/ssrn.4430515
D'Agostino, A., & Surico, P. (2012). A century of inflation forecasts. Review of Economics and Statistics, 94(4), 1097-1106.https://doi.org/10.1162/REST_a_00235
Devlin, J., Chang, M. W., Lee, K., y Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. En J. Burstein, C. Doran, y T. Solorio (Eds.), Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Vol. 1, pp. 4171-4186). Association for Computational Linguistics. https://doi.org/10.18653/v1/N19-1423
Diebold, F. X. & Mariano, R. S. (1995). Comparing predictive accuracy. Journal of Business and Economic Statistics, 13, 253-63.
https://doi.org/10.1080/07350015.1995.10524599
Faria e Castro, M., & Leibovici, F. (2024). Artificial Intelligence and Inflation Forecasts. Federal Reserve Bank of St. Louis Working Papers 2023-015.https://doi.org/10.20955/wp.2023.015
Faust, J., & Wright, J. H. (2013). Forecasting inflation. In G. Elliott & A. Timmermann (Eds.), Handbook of economic forecasting (Vol. 2, pp. 2-56). Elsevier.https://doi.org/10.1016/B978-0-444-53683-9.00001-3
Gao, J., Gebreegziabher, S. A., Choo, K. T. W., Li, T. J.-J., Perrault, S. T., & Malone, T. W. (2024). A Taxonomy for Human-LLM Interaction Modes: An Initial Exploration. Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, 1-11.
https://doi.org/10.1145/3613905.3650786
Google AI. (2023). Gemini: A Family of Highly Capable Multimodal Models. arXiv Preprint. doi:10.48550/arxiv.2312.11805.
Google AI. (2024). Google AI for Developers. Retrieved from https://ai.google.dev/pricing
Horton, J. J., Filippas, A., y Manning, B. S. (2024). Large language models as simulated economic agents: What can we learn from Homo silicus? (Working Paper No. 32013). National Bureau of Economic Research. https://doi.org/10.3386/w32013
Korinek, A. (2023). Language Models and Cognitive Automation for Economic Research. NBER Working Paper, 30957.
https://doi.org/10.3386/w30957
Lee, U. (2012). Forecasting Inflation for Inflation-Targeted Countries: A Comparison of the Predictive Performance of Alternative Inflation Forecasting Models. The Journal of Business and Economic Studies, 18(1), 75-95.
Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., & Dean, J. (2013). Distributed representations of words and phrases and their compositionality. Advances in Neural Information Processing Systems, 26.
OpenAI. (2023a). GPT-3.5: Technical Overview. Retrieved 18 September 2023, from https://platform.openai.com.
OpenAI. (2023b). GPT-4 Technical Report. Retrieved from https://openai.com/research/gpt-4.
OpenAI. (2024). GPT-4o Technical Report. Retrieved from https://openai.com/index/hello-gpt-4o/.
Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., & Sutskever, I. (2019). Language Models are Unsupervised Multitask Learners. Retrieved from https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf.
Shahriar, S., Lund, B. D., Mannuru, N. R., Arshad, M. A., Hayawi, K., Bevara, R. V. K., Mannuru, A., & Batool, L. (2024). Putting GPT-4o to the Sword: A Comprehensive Evaluation of Language, Vision, Speech, and Multimodal Proficiency. Applied Sciences, 14(17), 7782.
https://doi.org/10.3390/app14177782
Theil, H. (1967). Economics and Information Theory. North-Holland Publishing Company, Amsterdam.
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., Chi, E.H., Hashimoto, T., Vinyals, O., Liang, P., Dean, J., Fedus, W. (2022). Emergent Abilities of Large Language Models. arXiv preprint. arXiv:2206.07682.
Zhao, W.X., Zhou, K., Li, J., Tang, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., Du, Y., Yang, C., Chen, Y., Chen, Z., Jiang, J., Ren, R., Li, Y., Tang, X., Liu, Z., Liu, P., Nie, J., Wen, J. (2023). A Survey of Large Language Models. arXiv preprint. arXiv:2303.18223.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2024 Daniela Agostina Gonzalez
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Submission of manuscripts implies that the work described has not been published before (except in the form of an abstract or as part of thesis), that it is not under consideration for publication elsewhere and that, in case of acceptance, the authors agree to automatic transfer of the copyright to the Journal for its publication and dissemination. Authors retain the authors' right to use and share the article according to a personal or instutional use or scholarly sharing purposes; in addition, they retain patent, trademark and other intellectual property rights (including research data).
All the articles are published in the Journal under the Creative Commons license CC-BY-SA (Attribution-ShareAlike). It is allowed a commercial use of the work (always including the author attribution) and other derivative works, which must be released under the same license as the original work.
Up to Volume 21, this Journal has been licensing the articles under the Creative Commons license CC-BY-SA 3.0 ES. Starting from Volume 22, the Creative Commons license CC-BY-SA 4.0 is used.
