Statistical Identification and Backtesting of Mean-Reverting Forex Pairs Using Augmented Dickey-Fuller and Hurst Exponent Methods for Algorithmic Trading

Raniel Taripe

doi:10.64807/z5z11q86

Authors

Raniel B. Taripe Quezon City University Author
- Conceptualization
- Data Curation
- Formal Analysis
- Funding Acquisition
- Investigation
- Methodology
- Project Administration
- Resources
- Software
- Supervision
- Validation
- Visualization
- Writing – Original Draft Preparation
- Writing – Review & Editing

DOI:

https://doi.org/10.64807/z5z11q86

Abstract

This study investigates the mean-reverting behavior of major and minor foreign exchange (forex) currency pairs using statistical time-series analysis methods. The foreign exchange market is among the most liquid and widely traded financial markets globally, yet understanding the statistical dynamics of currency pairs remains a critical challenge for the development of algorithmic trading strategies. The primary objective of this research is to determine which major and minor forex pairs exhibit mean-reverting characteristics through the application of the Augmented Dickey-Fuller (ADF) test and Hurst exponent analysis, using ten years of daily closing price data spanning from 2016 to 2025. The study covers 28 currency pairs, comprising 7 major and 21 minor pairs. Exotic pairs were excluded due to lower liquidity and wider bid-ask spreads that may distort statistical results. Data was retrieved via the Google Finance function and analyzed using Python. The ADF test was employed to assess stationarity, while the Hurst exponent quantified the degree of mean-reversion or persistence in the return series. Results from the ADF test indicated rejection of the unit root null hypothesis for all currency pairs at the 5% significance level, confirming that the log return series are stationary. Hurst exponent analysis further revealed that all pairs exhibit values well below 0.5, consistent with mean-reverting behavior. The convergence of both statistical methods provides robust evidence that the log return dynamics of the analyzed forex pairs follow mean-reverting processes. These findings suggest opportunities for developing quantitative and algorithmic trading strategies based on statistical arbitrage principles. Future research is recommended to incorporate transaction costs, out-of-sample backtesting, exotic pairs, and additional statistical methods to validate further and expand these results.

Keywords:

Mean Reversion, Augmented Dickey-Fuller, Hurst Exponent, Forex Algorithmic Trading, Stationarity

Author Biography

Raniel B. Taripe, Quezon City University

Engr. Raniel B. Taripe is a multifaceted professional whose expertise spans quantitative analysis, data sciences, financial markets, and engineering analytics. He is a Certified Industrial Engineer, educator, and quantitative practitioner. He currently serves as a Faculty Member at Quezon City University under the College of Engineering – Industrial Engineering Department where he delivers sessions in data analytics and applied quantitative methods in industrial engineering such as operations research.

Engr. Raniel is highly fascinated in the interesting quantitative field of financial markets algorithmic trading. He holds three professional certifications in financial markets: Certified Financial Markets Professional (CFMP), Certified Advanced Technical Analyst (CATA), and Certified Advanced Equity Analyst (CAEA). He holds two master’s degrees: a Master in Business Administration major in Financial Management (MBA-FM) from Polytechnic University of the Philippines and a Master of Science in Financial Engineering (MScFE) from WorldQuant University in Washington, D.C., USA.

His graduate studies laid the foundation for his research and professional work in data analytics and the financial markets as his primary domain of expertise, exploring forex trading in his master’s thesis and algorithmic trading in his financial engineering capstone project. Building on this foundation, he currently mentors (by-invitation) aspiring forex traders at East West International Educational Specialists Inc., working with students and young professionals from various universities across the Philippines.

In academia, Engr. Raniel has mentored numerous Industrial Engineering student groups in capstone projects and feasibility studies, many of which have received awards and recognitions for excellence in research, innovation, and project implementation. His mentorship emphasizes rigorous quantitative analysis, real-world applicability, and data-driven decision-making.

With strong foundations in the quantitative pillars of Industrial Engineering—including Statistics, Operations Research, and Systems Optimization—Engr. Raniel has developed a growing research portfolio in data science and analytics. His primary research interest lies in Financial Data Science, an emerging interdisciplinary field at the intersection of Industrial Engineering, Financial Engineering, and computational analytics, where advanced statistical methods and machine learning techniques are applied to financial market data.

Continuing his pursuit of academic and professional growth, he is currently pursuing a Doctor of Philosophy in Industrial Engineering (PhD-IE) at University of the Philippines Diliman.

References

Bank for International Settlements. (2022). OTC foreign exchange turnover in April 2022 (Triennial Central Bank Survey). https://www.bis.org/statistics/rpfx22_fx.pdf

Bui, Q., & Ślepaczuk, R. (2022). Applying Hurst exponent in pair trading strategies on Nasdaq 100 index. Physica A: Statistical Mechanics and Its Applications, 592, 126784. https://doi.org/10.1016/j.physa.2021.126784

Enow, S. T. (2023). Investigating mean reversion in financial markets using Hurst model. International Journal of Research in Business and Social Science, 12(6), 197–201. https://doi.org/10.20525/ijrbs.v12i6.2664

Gatev, E., Goetzmann, W. N., & Rouwenhorst, K. G. (2006). Pairs trading: Performance of a relative-value arbitrage rule. The Review of Financial Studies, 19(3), 797–827. https://doi.org/10.1093/rfs/hhj020

Hasan, S., & Ghosh, R. (2024). Mean reversion dynamics in financial markets: Insights from Hurst exponent analysis. Indonesian Capital Market Review, 16(2), 86–99. https://doi.org/10.21002/icmr.v16i2.180

Kaczmarek, K., & Sevastjanov, P. (2023). A multi-model approach to the development of algorithmic trading systems for the Forex market. Expert Systems with Applications, 236, 121310. https://doi.org/10.1016/j.eswa.2023.121310

Kim, J., & Kim, H. (2022). Testing Long memory in the exchange rates and its Implications for the adaptive market hypothesis. Physica A: Statistical Mechanics and its Applications, 593, 126871. https://doi.org/10.1016/j.physa.2022.126871

Krauss, C. (2017). Statistical arbitrage pairs trading strategies: Review and outlook. Journal of Economic Surveys, 31(2), 513–545. https://doi.org/10.1111/joes.12153

López de Prado, M. (2018). Advances in financial machine learning. John Wiley & Sons. https://doi.org/10.1002/9781119482109

Rad, H., Low, R. K. Y., & Faff, R. (2016). The profitability of pairs trading strategies: Distance, cointegration, and copula methods. Quantitative Finance, 16(10), 1541–1558. https://doi.org/10.1080/14697688.2016.1164337

Ramos-Requena, J. P., Trinidad-Segovia, J. E., & Sánchez-Granero, M. Á. (2020). Some notes on the formation of a pair in pairs trading. Mathematics, 8(3), 348. https://doi.org/10.3390/math8030348

Requejo, D. (2024). Efficacy of a mean reversion trading strategy using True Strength Index [Working paper]. SSRN. https://doi.org/10.2139/ssrn.4708400

Sathyanarayana, S., & Mohanasundaram, T. (2024). Stationarity and unit roots in time series. Research Advances in Journal of Management and Social Sciences, 4(1), 50–65. https://doi.org/10.54060/RAJMSS/004.01.007

Silva, J., & Ferreira, P. (2023). Testing efficiency in cryptocurrency markets: A detrended fluctuation analysis approach. Research in International Business and Finance, 64, 101835. https://doi.org/10.1016/j.ribaf.2022.101835

Vogl, M. (2023). Hurst exponent dynamics of S&P 500 returns: Implications for market efficiency, long memory, multifractality and financial crises predictability by application of a nonlinear dynamics analysis framework. Chaos, Solitons & Fractals, 166, 112884. https://doi.org/10.1016/j.chaos.2022.112884

Wang, H., & Li, M. (2022). Testing long memory and structural changes in financial time series using a time-varying Hurst exponent. Physica A: Statistical Mechanics and its Applications, 604, 127914. https://doi.org/10.1016/j.physa.2022.127914

Zhang, Z. (2023). Comparison of the performances for the mean-reversion strategy for exchange rate. BCP Business & Management, 38, 1811–1817. https://doi.org/10.54691/bcpbm.v38i.3971