Personalized dynamic pricing strategies with fairness in the credit market

Haiying Liu; Chao Wu; Wenjie Bi; Haiying Liu; Chao Wu; Wenjie Bi

doi:10.3934/QFE.2026019

Quantitative Finance and Economics

2026, Volume 10, Issue 3: 481-503. doi: 10.3934/QFE.2026019

Previous Article Next Article

Research article

Personalized dynamic pricing strategies with fairness in the credit market

1.
Accounting School, Hunan University of Finance and Economics, Changsha, China
2.
Institute of Intelligent Finance, Hunan University of Finance and Economics, Changsha, China
3.
School of Business Administration, South China University of Technology, Guangzhou, China
4.
Business School, Central South University, Changsha, China

Received: 24 December 2025 Revised: 31 March 2026 Accepted: 02 June 2026 Published: 16 July 2026
JEL Codes: C44, D81, G12

With the rapid development of the credit market, financial institutions are increasingly focusing on pricing strategies for credit products. Traditional pricing models are mainly based on economic frameworks that are often inadequate to adapt to complex and dynamic market environments, resulting in inefficiencies and lost revenue. Advances in information technology enable financial institutions to accurately predict customer characteristics, thereby facilitating personalized pricing. While personalized pricing can increase revenue, it also raises consumer concerns about fairness. Therefore, this study introduces the concept of representative fairness to strike a balance between institutional profitability and consumer perceptions of fairness in personalized scenarios. Personalized pricing strategies are developed by integrating a representative fairness clustering algorithm with a multi-armed bandit (MAB) algorithm. We propose the K-means-Thompson Sampling (KM-TS) and Representative Fairness K-means-Thompson Sampling (RFKM-TS) algorithms and conduct numerical experiments using a virtual bank dataset from Ireland. The experimental results show that the KM-TS algorithm outperforms real pricing and other MAB algorithms. When representative fairness is incorporated, the RFKM-TS algorithm continues to generate higher revenues than actual pricing, despite a slight performance degradation. The proposed algorithm effectively addresses data sparsity and provides a valuable reference for pricing strategies in personalized decision scenarios.
- credit market,
- MAB algorithm,
- personalized pricing strategies,
- KM-TS algorithm,
- RFKM-TS algorithm
Citation: Haiying Liu, Chao Wu, Wenjie Bi. Personalized dynamic pricing strategies with fairness in the credit market[J]. Quantitative Finance and Economics, 2026, 10(3): 481-503. doi: 10.3934/QFE.2026019

Related Papers:

Abstract

With the rapid development of the credit market, financial institutions are increasingly focusing on pricing strategies for credit products. Traditional pricing models are mainly based on economic frameworks that are often inadequate to adapt to complex and dynamic market environments, resulting in inefficiencies and lost revenue. Advances in information technology enable financial institutions to accurately predict customer characteristics, thereby facilitating personalized pricing. While personalized pricing can increase revenue, it also raises consumer concerns about fairness. Therefore, this study introduces the concept of representative fairness to strike a balance between institutional profitability and consumer perceptions of fairness in personalized scenarios. Personalized pricing strategies are developed by integrating a representative fairness clustering algorithm with a multi-armed bandit (MAB) algorithm. We propose the K-means-Thompson Sampling (KM-TS) and Representative Fairness K-means-Thompson Sampling (RFKM-TS) algorithms and conduct numerical experiments using a virtual bank dataset from Ireland. The experimental results show that the KM-TS algorithm outperforms real pricing and other MAB algorithms. When representative fairness is incorporated, the RFKM-TS algorithm continues to generate higher revenues than actual pricing, despite a slight performance degradation. The proposed algorithm effectively addresses data sparsity and provides a valuable reference for pricing strategies in personalized decision scenarios.

References

[1]	Arneson RJ (2004) Luck egalitarianism interpreted and defended. Philos Topics 32: 1–20. https://doi.org/10.5840/philtopics2004321/217 doi: 10.5840/philtopics2004321/217
[2]	Aslan A, Bakir I, Vis I (2020) A dynamic thompson sampling hyper-heuristic framework for learning activity planning in personalized learning. Eur J Oper Res 286: 673–688. https://doi.org/10.1016/j.ejor.2020.03.038 doi: 10.1016/j.ejor.2020.03.038
[3]	Bayraktar E, Yang B (2011) A unified framework for pricing credit and equity derivatives. Math Financ: Int J Math Stat Financ Econ 21: 493–517. https://doi.org/10.1111/j.1467-9965.2010.00435.x doi: 10.1111/j.1467-9965.2010.00435.x
[4]	Bélanger A, Shreve SE, Wong D (2004) A general framework for pricing credit risk. Math Financ: Int J Math Stat Financ Econ 14: 317–350. https://doi.org/10.1111/j.0960-1627.2004.t01-1-00193.x doi: 10.1111/j.0960-1627.2004.t01-1-00193.x
[5]	Bernstein F, Modaresi S, Sauré D (2019) A dynamic clustering approach to data-driven assortment personalization. Manage Sci 65: 2095–2115. https://doi.org/10.1287/mnsc.2018.3031 doi: 10.1287/mnsc.2018.3031
[6]	Burgard JP, Moreira Costa C, Schmidt M (2024) Robustification of the k-means clustering problem and tailored decomposition methods: when more conservative means more accurate. Ann Oper Res 339: 1525–1568. https://doi.org/10.1007/s10479-022-04818-w doi: 10.1007/s10479-022-04818-w
[7]	Chen R, Zhou H, Jin C, et al. (2020) Discount or premium? Pricing of structured products: An analysis of Chinese market. Int Rev Financ Anal 70: 101493. https://doi.org/10.1016/j.irfa.2020.101493 doi: 10.1016/j.irfa.2020.101493
[8]	Chi G, Dong B, Zhou Y, et al. (2024) Long-horizon predictions of credit default with inconsistent customers. Technol Forecast Soc 198: 123008. https://doi.org/10.1016/j.techfore.2023.123008 doi: 10.1016/j.techfore.2023.123008
[9]	Chun SY, Lejeune MA (2020) Risk-Based Loan Pricing: Portfolio Optimization Approach with Marginal Risk Contribution. Manage Sci 66: 3735–3753. https://doi.org/10.1287/mnsc.2019.3378 doi: 10.1287/mnsc.2019.3378
[10]	Corvino R, Fusai G (2022) Default risk premium and asset prices. J Financ Stabil 60: 101014. https://doi.org/10.1016/j.jfs.2022.101014 doi: 10.1016/j.jfs.2022.101014
[11]	Deepak P, Abraham SS (2020) Representativity fairness in clustering. In: PROCEEDINGS of the 12th ACM conference on Web Science, 202–211. https://doi.org/10.1145/3394231.3397910
[12]	Emekter R, Tu Y, Jirasakuldech B, et al. (2015) Evaluating credit risk and loan performance in online Peer-to-Peer (P2P) lending. Appl Econ 47: 54–70. https://doi.org/10.1080/00036846.2014.962222 doi: 10.1080/00036846.2014.962222
[13]	Ewald CO, Wu Y, Zhang A (2023) Pricing Asian options with stochastic convenience yield and jumps. Quant Financ 23: 677–692. https://doi.org/10.1080/14697688.2022.2160799 doi: 10.1080/14697688.2022.2160799
[14]	Ferreira KJ, Simchi-Levi D, Wang H (2018) Online network revenue management using Thompson sampling. Oper Res 66: 1586–1602. https://doi.org/10.1287/opre.2018.1755 doi: 10.1287/opre.2018.1755
[15]	Filipe SF, Grammatikos T, Michala D (2016) Pricing default risk: The good, the bad, and the anomaly. J Financ Stabil 26: 190–213. https://doi.org/10.1016/j.jfs.2016.07.001 doi: 10.1016/j.jfs.2016.07.001
[16]	Ganong P, Noel P (2020) Liquidity versus wealth in household debt obligations: Evidence from housing policy in the great recession. Am Econ Rev 110: 3100–3138. https://doi.org/10.1257/aer.20181243 doi: 10.1257/aer.20181243
[17]	Ganti R, Sustik M, Tran Q, et al. (2018) Thompson sampling for dynamic pricing. arXiv preprint arXiv:1802.03050.
[18]	Gürsoy K (2024) Multi-armed bandit games. Ann Oper Res. https://doi.org/10.1007/s10479-024-06336-3 doi: 10.1007/s10479-024-06336-3
[19]	Han K, He Y, Liu AX, et al. (2020) Differentially private and budget-limited bandit learning over matroids. INFORMS J Comput 32: 790–804. https://doi.org/10.1287/ijoc.2019.0903 doi: 10.1287/ijoc.2019.0903
[20]	Hu B, Hu YP (2024) A pricing model system for small and micro loan insurance considering limited claims Int Rev Financ Anal 93: 103212. https://doi.org/10.1016/j.irfa.2024.103212 doi: 10.1016/j.irfa.2024.103212
[21]	Im S, Moseley B, Munagala K, et al. (2020) Dynamic weighted fairness with minimal disruptions. Proceedings of the ACM on Measurement and Analysis of Computing Systems 4: 1–18. https://doi.org/10.1145/3379485 doi: 10.1145/3379485
[22]	Jiang Y, Ji X, Wu J, et al. (2023) Behavior-based pricing and consumer fairness concerns with green product design. Ann Oper Res. https://doi.org/10.1007/s10479-023-05198-5 doi: 10.1007/s10479-023-05198-5
[23]	Kozodoi N, Jacob J, Lessmann S (2022) Fairness in credit scoring: Assessment, implementation and profit implications. Eur J Oper Res 297: 1083–1094. https://doi.org/10.1016/j.ejor.2021.06.023 doi: 10.1016/j.ejor.2021.06.023
[24]	Kumar IE, Hines KE, Dickerson JP (2022) Equalizing credit opportunity in algorithms: Aligning algorithmic fairness research with us fair lending regulation. In: Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society, 357–368. https://doi.org/10.1145/3514094.3534154
[25]	Letard A, Gutowski N, Camp O, et al. (2024) Bandit algorithms: A comprehensive review and their dynamic selection from a portfolio for multicriteria top-k recommendation. Expert Syst Appl, 123151. https://doi.org/10.1016/j.eswa.2024.123151 doi: 10.1016/j.eswa.2024.123151
[26]	Li L, Chu W, Langford J, et al. (2010) A contextual-bandit approach to personalized news article recommendation. In: Proceedings of the 19th international conference on World wide web, 661–670. https://doi.org/10.1145/1772690.1772758
[27]	Liu J, Zhao H, Liu S (2018) A study of service pricing with unfairness averse customers. Chinese J Manage Sci 26: 46. Available from: https://www.zgglkx.com/CN/10.16381/j.cnki.issn1003-207x.2018.02.005.
[28]	Louhichi A, Boujelbene Y (2020) Credit risk pricing and the rationality of lending decision-making within dual banking systems: A parametric approach. Econ Syst 44: 100740. https://doi.org/10.1016/j.ecosys.2019.100740 doi: 10.1016/j.ecosys.2019.100740
[29]	MacQueen J (1967) Some methods for classification and analysis of multivariate observations. Univ California Press. Available from: https://cir.nii.ac.jp/crid/1570854176032593536.
[30]	Marzban S, Delage E, Li JY (2022) Equal risk pricing and hedging of financial derivatives with convex risk measures. Quant Financ 22: 47–73. https://doi.org/10.1080/14697688.2021.1993614 doi: 10.1080/14697688.2021.1993614
[31]	Miao S, Chen X, Chao X, et al. (2022) Context-based dynamic pricing with online clustering. Prod Oper Manag 31: 3559–3575. https://doi.org/10.1111/poms.13783 doi: 10.1111/poms.13783
[32]	Misra K, Schwartz EM, Abernethy J (2019) Dynamic online pricing with incomplete information using multiarmed bandit experiments. Mark Sci 38: 226–252. https://doi.org/10.1287/mksc.2018.1129 doi: 10.1287/mksc.2018.1129
[33]	Niu Z, Huang P, Wang J (2017) Online-offline uniform pricing strategy of multi-channel retailer based on fairness preferences. Chinese J Manage Sci 25: 147. Available from: https://www.zgglkx.com/CN/10.16381/j.cnki.issn1003-207x.2017.03.017.
[34]	Rawls J (1971) A theory of justice. Cambridge, MA, 1–9. https://doi.org/10.4159/9780674042605
[35]	Rodgers W, Hudson R, Economou F (2023) Modelling credit and investment decisions based on AI algorithmic behavioral pathways. Technol Forecast Soc 191: 122471. https://doi.org/10.1016/j.techfore.2023.122471 doi: 10.1016/j.techfore.2023.122471
[36]	Schwartz EM, Bradlow ET, Fader PS (2017) Customer acquisition via display advertising using multi-armed bandit experiments. Market Sci 36: 500–522. https://doi.org/10.1287/mksc.2016.1023 doi: 10.1287/mksc.2016.1023
[37]	Silva N, Werneck H, Silva T, et al. (2022) Multi-armed bandits in recommendation systems: A survey of the state-of-the-art and future directions. Expert Syst Appl 197: 116669. https://doi.org/10.1016/j.eswa.2022.116669 doi: 10.1016/j.eswa.2022.116669
[38]	Sun W, Zhang X, Li M, et al. (2023) Interpretable high-stakes decision support system for credit default forecasting. Technol Forecast Soc 196: 122825. https://doi.org/10.1016/j.techfore.2023.122825 doi: 10.1016/j.techfore.2023.122825
[39]	Trovo F, Paladino S, Restelli M, et al. (2015) Multi-armed bandit for pricing. In: Proc 12th European Workshop on Reinforcement Learning, 1–9. Available from: https://re.public.polimi.it/bitstream/11311/1151563/2/trovo2015multiarmed.pdf [Accessed on 15 Sep 2024].
[40]	Zhang S (2023) Takeout Platform Uses Big Data for Price Discrimination–-Meituan Takeout Platform Case Study. J Res Soc Sci Hum 2: 36–41. https://doi.org/10.56397/JRSSH.2023.03.05 doi: 10.56397/JRSSH.2023.03.05
[41]	Ziko IM, Yuan J, Granger E, et al. (2021) Variational fair clustering. In: Proceedings of the AAAI Conference on Artificial Intelligence 35: 11202–11209. https://doi.org/10.1609/aaai.v35i12.17336

Reader Comments

Your name:*

Email:*
© 2026 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)