State-Dependent Asset Pricing and Machine-Learning Return Forecasting: Evidence from the Pakistan Stock Exchange
DOI:
https://doi.org/10.62843/jssr.v6i2.765Keywords:
Asset Pricing, Machine Learning, Random Forest, Artificial Neural Network, Adaptive Market Hypothesis, Fama-French Five-Factor Model, Arbitrage Pricing Theory, Emerging Markets, Pakistan Stock Exchange, Return PredictabilityAbstract
This study asks whether asset-pricing relationships and return predictability in an emerging market vary across market conditions, and whether machine-learning methods capture this state dependence more flexibly than linear approaches. Using 18 Pakistan Stock Exchange (PSX) portfolio-return series (February 2010-December 2025, N=191), we estimate Fama-French five-factor (FF5) and Arbitrage Pricing Theory (APT)-motivated macroeconomic specifications with ordinary least squares (OLS), an artificial neural network (ANN) and Random Forest (RF), separately for the full sample and for rolling-window-classified Bullish (n=26), Bearish (n=30) and Neutral (n=124) market conditions. The FF5/OLS specification has substantial in-sample explanatory power (average R²=0.397); the macroeconomic/OLS specification is statistically significant for several variables but has much lower explanatory power (average R²=0.061). On a 70/30 chronological holdout, Random Forest extracts considerable nonlinear predictive structure from the FF5 factors (average out-of-sample R²=0.645; mean directional accuracy=80.3%), while the ANN is more sensitive to the limited sample (R²=0.055). Macroeconomic machine-learning models do not generalise out of sample (negative average R²). Factor and feature importance shift materially across market conditions. These patterns are consistent with state-dependent asset pricing and the Adaptive Market Hypothesis, but the evidence is conditional on the sample, predictor set and model specification, and does not establish universal machine-learning superiority or investment profitability.
References
Armano, G., Marchesi, M., & Murru, A. (2005). A hybrid genetic-neural architecture for stock indexes forecasting. Information Sciences, 170(1), 3–33. https://doi.org/10.1016/j.ins.2003.03.023
Bagnara, M., Goodell, J. W., Ortmann, R., & Staudinger, M. (2024). Asset pricing and machine learning: A critical review. Journal of Economic Surveys, 38(1), 27–56. https://doi.org/10.1111/joes.12532
Bradrania, R., & Pirayesh Neghab, D. (2022). State-dependent asset allocation using neural networks. The European Journal of Finance, 28(11), 1125–1150. https://doi.org/10.1080/1351847X.2021.1960404
Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32. https://doi.org/10.1023/A:1010933404324
Cao, Q., Parry, M. E., & Leggio, K. B. (2011). The three-factor model and artificial neural networks: Predicting stock price movement in China. Annals of Operations Research, 185(1), 25–44. https://doi.org/10.1007/s10479-009-0618-0
Carvalhal, A., & Ribeiro, T. (2008). Do artificial neural networks provide better forecasts? Evidence from Latin American stock indexes. Latin American Business Review, 9(3), 92–110. https://doi.org/10.1080/10978520802035463
Chiah, M., Chai, D., Zhong, A., & Li, S. (2016). A better model? An empirical investigation of the Fama–French five-factor model in Australia. International Review of Finance, 16(4), 595–638. https://doi.org/10.1111/irfi.12099
Chen, L., Pelger, M., & Zhu, J. (2024). Deep learning in asset pricing. Management Science, 70(2), 714–750. https://doi.org/10.1287/mnsc.2023.4695
Chen, S.-J., & Jordan, B. D. (1993). Some empirical tests in the arbitrage pricing theory: Macro variables vs. derived factors. Journal of Banking & Finance, 17(1), 65–89. https://doi.org/10.1016/0378-4266(93)90080-W
Cheng, T., Jiang, S., Zhao, A. B., & Zhao, J. (2025). Is machine learning a necessity? A regression-based approach for stock return prediction. Journal of Empirical Finance, 81, Article 101598. https://doi.org/10.1016/j.jempfin.2025.101598
Connor, G., & Korajczyk, R. A. (1995). The arbitrage pricing theory and multifactor models of asset returns. In R. Jarrow, V. Maksimovic, & W. T. Ziemba (Eds.), Handbooks in operations research and management science: Finance (Vol. 9, pp. 87–144). Elsevier. https://doi.org/10.1016/S0927-0507(05)80048-9
Fallahgoul, H., Franstianto, V., & Lin, X. (2024). Asset pricing with neural networks: Significance tests. Journal of Econometrics, 238(1), Article 105574. https://doi.org/10.1016/j.jeconom.2023.105574
Fama, E. F., & French, K. R. (2017). International tests of a five-factor asset pricing model. Journal of Financial Economics, 123(3), 441–463. https://doi.org/10.1016/j.jfineco.2016.11.004
Fang, M., & Taylor, S. (2021). A machine learning based asset pricing factor model comparison on anomaly portfolios. Economics Letters, 204, Article 109919. https://doi.org/10.1016/j.econlet.2021.109919
Fatima, S., & Hussain, G. (2008). Statistical models of KSE100 index using hybrid financial systems. Neurocomputing, 71(13–15), 2742–2746. https://doi.org/10.1016/j.neucom.2007.11.044
Goyal, A., & Welch, I. (2008). A comprehensive look at the empirical performance of equity premium prediction. The Review of Financial Studies, 21(4), 1455–1508. https://doi.org/10.1093/rfs/hhm014
Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223–2273. https://doi.org/10.1093/rfs/hhaa009
Hanauer, M. X., & Kalsbach, T. (2023). Machine learning and the cross-section of emerging market stock returns. Emerging Markets Review, 55, Article 101022. https://doi.org/10.1016/j.ememar.2023.101022
Haqqani, K., & Aleem, M. (2020). Testing liquidity augmented Fama-French five-factor model in Pakistan Stock Exchange. Global Economics Review, 5(1), 255–265. https://doi.org/10.31703/ger.2020(V-I).21
He, Z., Zhu, J., & Zhu, X. (2015a). Dynamic factors and asset pricing: International and further U.S. evidence. Pacific-Basin Finance Journal, 32, 21–39. https://doi.org/10.1016/j.pacfin.2015.02.002
He, Z., Zhu, J., & Zhu, X. (2015b). Multi-factor volatility and stock returns. Journal of Banking & Finance, 61(S2), S132–S149. https://doi.org/10.1016/j.jbankfin.2015.09.013
Idowu, P. A., Osakwe, C., Kayode, A. A., & Adagunodo, E. R. (2012). Prediction of stock market in Nigeria using artificial neural network. International Journal of Intelligent Systems and Applications, 4(11), 68–74. https://doi.org/10.5815/ijisa.2012.11.08
Jan, M. N. (2019). Asset pricing and artificial neural networks: A case of Pakistan’s equity market [Doctoral dissertation, Capital University of Science and Technology].
Kazak, H., Kumar, S., Gündüz, M. A., & Akcan, A. T. (2025). Metaheuristic-optimized ANFIS and ANN models for stock price forecasting: Evidence from the Borsa Istanbul 100 index. Discover Artificial Intelligence, 5, Article 272. https://doi.org/10.1007/s44163-025-00395-6
Kelly, B. T., Malamud, S., & Zhou, K. (2024). The virtue of complexity in return prediction. The Journal of Finance, 79(1), 459–503. https://doi.org/10.1111/jofi.13298
Khan, N., Zada, H., Ahmed, S., Shah, F. A., & Jan, S. (2023). Human capital-based four-factor asset pricing model: An empirical study from Pakistan. Heliyon, 9(5), Article e16328. https://doi.org/10.1016/j.heliyon.2023.e16328
Liang, D., Lu, C.-C., Tsai, C.-F., & Shih, G.-A. (2024). Forecasting corporate financial performance with deep learning and interpretable ALE method: Evidence from China. Journal of Forecasting, 43(7), 2540–2571. https://doi.org/10.1002/for.3138
Lo, A. W. (2004). The adaptive markets hypothesis: Market efficiency from an evolutionary perspective. The Journal of Portfolio Management, 30(5), 15–29. https://doi.org/10.3905/jpm.2004.442611
Munir, T., Saeed, M., & Aziz, B. (2023). Forecasting the Pakistan’s leading stock exchange during COVID-19 using machine learning. Research Square. https://doi.org/10.21203/rs.3.rs-2498599/v1
Nyberg, H. (2013). Predicting bear and bull stock markets with dynamic binary probit models. Journal of Banking & Finance, 37(8), 2952–2963. https://doi.org/10.1016/j.jbankfin.2013.05.008
Priestley, R. (1996). The arbitrage pricing theory, macroeconomic and financial factors, and expectations generating processes. Journal of Banking & Finance, 20(5), 869–890. https://doi.org/10.1016/0378-4266(95)00035-6
Raza, H., & Akhtar, Z. (2024). Predicting stock prices in the Pakistan market using machine learning and technical indicators. Modern Finance, 2(2), 46–63. https://doi.org/10.61351/mf.v2i2.167
Ross, S. A. (1976). The arbitrage theory of capital asset pricing. Journal of Economic Theory, 13(3), 341–360. https://doi.org/10.1016/0022-0531(76)90046-6
Sha, Y., & Gao, R. (2019). Which is the best: A comparison of asset pricing factor models in Chinese mutual fund industry. Economic Modelling, 83, 8–16. https://doi.org/10.1016/j.econmod.2019.09.016
Ullah, R., Jan, M. N., & Tahir, M. (2025). Unveiling the optimal factor model in Pakistan: A machine learning approach using support vector regression and extreme gradient boosting algorithms. Future Business Journal, 11, Article 137. https://doi.org/10.1186/s43093-025-00560-4
Yao, H., Xia, S., & Liu, H. (2022). Six-factor asset pricing and portfolio investment via deep learning: Evidence from Chinese stock market. Pacific-Basin Finance Journal, 76, Article 101886. https://doi.org/10.1016/j.pacfin.2022.101886
Zaremba, A., & Szczygielski, J. J. (2019). Limits to arbitrage, investor sentiment, and factor returns in international government bond markets. Economic Research-Ekonomska Istraživanja, 32(1), 1727–1743. https://doi.org/10.1080/1331677X.2019.1638286
Zegadło, P. (2022). Identifying bull and bear market regimes with a robust rule-based method. Research in International Business and Finance, 60, Article 101603. https://doi.org/10.1016/j.ribaf.2021.101603
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Copyright in the Journal of Social Sciences Review is retained by the author(s). Authors also grant any third party the right to use the article freely as long as its integrity is maintained and its original authors, citation details and publisher are identified.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
JSSR's Editorial Board shares the vision of providing free access to information, education, and science for everyone, thus promoting its content through an OPEN ACCESS POLICY, fulfilling the DOAJ definition of open access. The JSSR adheres to an Open Access and Copyright Licensing Policy based on the belief that making research freely accessible to the public promotes greater global knowledge sharing.
The JSSR uses the Creative Commons Attribution-NonCommercial 4.0 International License. The authors who apply and publish in JSSR consent to abide by the copyright policy set out in the Creative Commons 4.0 license (Attribution-NonCommercial 4.0 International license).
- Copyright in the Journal of Social Sciences Review is retained by the author(s).
- Authors also grant any third party the right to use the article freely as long as its integrity is maintained and its original authors, citation details and publisher are identified.
While "By 'open access' to this literature, we mean its free availability on the public internet, permitting any users to read, download, copy, distribute, print, search, or link to the full texts of these articles, crawl them for indexing, pass them as data to software, or use them for any other lawful purpose, without financial, legal, or technical barriers other than those inseparable from gaining access to the internet itself."
