Neural ARFIMA model for forecasting BRIC exchange rates with long memory under oil shocks and policy uncertainties

2025-09-08

AI Summary

BRIC exchange rates exhibit long memory, nonlinearity, and nonstationarity, and are subject to multiple external shocks—including global economic policy uncertainty, U.S. equity volatility, oil price fluctuations, and interest rate differentials—posing significant modeling challenges. To address these, this paper proposes the Neural Autoregressive Fractionally Integrated Moving Average (NARFIMA) model. NARFIMA is the first to integrate ARFIMA’s capacity for capturing long-memory dynamics with neural networks’ ability to model nonlinear dependencies; its asymptotic stationarity is rigorously established via Markov chain theory. Furthermore, conformal prediction is incorporated to construct statistically reliable prediction intervals. Empirical evaluation across six forecasting horizons demonstrates that NARFIMA significantly outperforms conventional statistical models (e.g., ARFIMA, VAR) and state-of-the-art machine learning methods (e.g., LSTM, XGBoost), markedly improving forecast accuracy for BRIC exchange rates. This advances tools for exchange rate risk management and macroeconomic policymaking in emerging markets.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Other Foundations of Machine Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Accurate forecasting of exchange rates remains a persistent challenge, particularly for emerging economies such as Brazil, Russia, India, and China (BRIC). These series exhibit long memory, nonlinearity, and non-stationarity properties that conventional time series models struggle to capture. Additionally, there exist several key drivers of exchange rate dynamics, including global economic policy uncertainty, US equity market volatility, US monetary policy uncertainty, oil price growth rates, and country-specific short-term interest rate differentials. These empirical complexities underscore the need for a flexible modeling framework that can jointly accommodate long memory, nonlinearity, and the influence of external drivers. To address these challenges, we propose a Neural AutoRegressive Fractionally Integrated Moving Average (NARFIMA) model that combines the long-memory representation of ARFIMA with the nonlinear learning capacity of neural networks, while flexibly incorporating exogenous causal variables. We establish theoretical properties of the model, including asymptotic stationarity of the NARFIMA process using Markov chains and nonlinear time series techniques. We quantify forecast uncertainty using conformal prediction intervals within the NARFIMA framework. Empirical results across six forecast horizons show that NARFIMA consistently outperforms various state-of-the-art statistical and machine learning models in forecasting BRIC exchange rates. These findings provide new insights for policymakers and market participants navigating volatile financial conditions. The exttt{narfima} extbf{R} package provides an implementation of our approach.

Problem

• Forecasting BRIC exchange rates with long memory properties
• Incorporating external drivers like oil shocks and policy uncertainties
• Addressing nonlinearity and non-stationarity in emerging market currencies

Innovation

• Combines ARFIMA long-memory with neural network nonlinear learning
• Incorporates exogenous variables like oil prices and policy uncertainties
• Uses conformal prediction for quantifying forecast uncertainty intervals