Quantitative Finance

Yoo, Jiwook

doi: 10.48550/arxiv.2509.08096pmid: N/A

Abstract:This article proposes a calibration framework for complex option pricing models that jointly fits market option prices and the term structure of variance. Calibrated models under the conventional objective function, the sum of squared errors in Black-Scholes implied volatilities, can produce model-implied variance term structures with large errors relative to those observed in the market and implied by option prices. I show that this can occur even when the model-implied volatility surface closely matches the volatility surface observed in the market. The proposed joint calibration addresses this issue by augmenting the conventional objective function with a penalty term for large deviations from the observed variance term structure. This augmented objective function features a hyperparameter that governs the relative weight placed on the volatility surface and the variance term structure. I test this framework on a jump-diffusion model with stochastic volatility in two calibration exercises: the first using volatility surfaces generated under a Bates model, and the second using a panel of S&P 500 equity index options covering the 1996-2023 period. I demonstrate that the proposed method is able to fit observed option prices well while delivering realistic term structures of variance. Finally, I provide guidance on the choice of hyperparameters based on the results of these numerical exercises.

Sharma, Chandradew

doi: 10.48550/arxiv.2509.00697pmid: N/A

Abstract:This study presents a unified, distribution-aware, and complexity-informed framework for understanding equity return dynamics in the Indian market, using 34 years (1990 to 2024) of Nifty 50 index data. Addressing a key gap in the literature, we demonstrate that the price to earnings ratio, as a valuation metric, may probabilistically map return distributions across investment horizons spanning from days to decades. Return profiles exhibit strong asymmetry. One-year returns show a 74 percent probability of gain, with a modal return of 10.67 percent and a reward-to-risk ratio exceeding 5. Over long horizons, modal CAGRs surpass 13 percent, while worst-case returns remain negative for up to ten years, defining a historical trapping period. This horizon shortens to six years in the post-1999 period, reflecting growing market resilience. Conditional analysis of the P/E ratio reveals regime-dependent outcomes. Low valuations (P/E less than 13) historically show zero probability of loss across all horizons, while high valuations (P/E greater than 27) correspond to unstable returns and extended breakeven periods. To uncover deeper structure, we apply tools from complexity science. Entropy, Hurst exponents, and Lyapunov indicators reveal weak persistence, long memory, and low-dimensional chaos. Information-theoretic metrics, including mutual information and transfer entropy, confirm a directional and predictive influence of valuation on future returns. These findings offer actionable insights for asset allocation, downside risk management, and long-term investment strategy in emerging markets. Our framework bridges valuation, conditional distributions, and nonlinear dynamics in a rigorous and practically relevant manner.

Karasan, Abdullah; Alp, Özge Sezgin

doi: 10.48550/arxiv.2509.01156pmid: N/A

Abstract:Filtering signal from noise is fundamental to accurately assessing spillover effects in financial markets. This study investigates denoised return and volatility spillovers across a diversified set of markets, spanning developed and developing economies as well as key asset classes, using a neural network-based denoising architecture. By applying denoising to the covariance matrices prior to spillover estimation, we disentangle signal from noise. Our analysis covers the period from late 2014 to mid-2025 and adopts both static and time-varying frameworks. The results reveal that developed markets predominantly serve as net transmitters of volatility spillovers under normal conditions, but often transition into net receivers during episodes of systemic stress, such as the Covid-19 pandemic. In contrast, developing markets display heightened instability in their spillover roles, frequently oscillating between transmitter and receiver positions. Denoising not only clarifies the dynamic and heterogeneous nature of spillover channels, but also sharpens the alignment between observed spillover patterns and known financial events. These findings highlight the necessity of denoising in spillover analysis for effective monitoring of systemic risk and market interconnectedness.

Zhuang, Jirong; Wu, Xuan

doi: 10.48550/arxiv.2509.11928pmid: N/A

Abstract:We treat implied volatility surface (IVS) reconstruction as a learning problem guided by two principles. First, we adopt a meta-learning view that trains across trading days to learn a procedure that maps sparse option quotes to a full IVS via conditional prediction, avoiding per-day calibration at test time. Second, we impose a structural prior via transfer learning: pre-train on SABR-generated dataset to encode geometric prior, then fine-tune on historical market dataset to align with empirical patterns. We implement both principles in a single attention-based Neural Process (Volatility Neural Process, VolNP) that produces a complete IVS from a sparse context set in one forward pass. On SPX options, the VolNP outperforms SABR, SSVI, and Gaussian process. Relative to an ablation trained only on market data, the SABR-induced prior reduces RMSE by about 40% and suppresses large errors, with pronounced gains at long maturities where quotes are sparse. The resulting model is fast (single pass), stable (no daily recalibration), and practical for deployment at scale.

Sester, Julian; Xu, Huansang

doi: 10.48550/arxiv.2509.25899pmid: N/A

Abstract:In this paper, we propose an alternative valuation approach for CAT bonds where a pricing formula is learned by deep neural networks. Once trained, these networks can be used to price CAT bonds as a function of inputs that reflect both the current market conditions and the specific features of the contract. This approach offers two main advantages. First, due to the expressive power of neural networks, the trained model enables fast and accurate evaluation of CAT bond prices. Second, because of its fast execution the trained neural network can be easily analyzed to study its sensitivities w.r.t. changes of the underlying market conditions offering valuable insights for risk management.

Cantarella, Michele; Maurizio, Maria Cristina; Serti, Francesco

doi: 10.48550/arxiv.2509.10193pmid: N/A

Abstract:This paper evaluates the short and medium-term effectiveness of payroll tax reductions aimed at promoting the permanent conversion of temporary contracts through social contribution exemptions. Using rich administrative data from Tuscany, providing detailed employment histories, we exploit a unique change in eligibility criteria in 2018 to estimate the causal impact of these exemptions. We find that the incentives immediately increased the probability of conversion, with no evidence of substitution against non-eligible cohorts. However, these positive effects were short-lived and appear to reflect anticipated conversions. Indeed, in the medium term, we find no persistent effects on a broad set of employment outcomes -- including whether the worker remains in the same permanent job, holds any permanent position, continues working in the same firm or sector, and how long has kept working -- and no evidence of heterogeneous effects across firm or worker characteristics.

Yaqoob, Ahad; Abdullah, Syed M.

doi: 10.48550/arxiv.2509.14401pmid: N/A

Abstract:The application of deep learning models for stock price forecasting in emerging markets remains underexplored despite their potential to capture complex temporal dependencies. This study develops and evaluates a Long Short-Term Memory (LSTM) network model for predicting the closing prices of ten major stocks across diverse sectors of the Pakistan Stock Exchange (PSX). Utilizing historical OHLCV data and an extensive set of engineered technical indicators, we trained and validated the model on a multi-year dataset. Our results demonstrate strong predictive performance ($R^2 > 0.87$) for stocks in stable, high-liquidity sectors such as power generation, cement, and fertilizers. Conversely, stocks characterized by high volatility, low liquidity, or sensitivity to external shocks (e.g., global oil prices) presented significant forecasting challenges. The study provides a replicable framework for LSTM-based forecasting in data-scarce emerging markets and discusses implications for investors and future research.

Tepelyan, Ruslan

doi: 10.48550/arxiv.2509.16137pmid: N/A

Abstract:OHLC bar data is a widely used format for representing financial asset prices over time due to its balance of simplicity and informativeness. Bloomberg has recently introduced a new bar data product that includes additional timing information-specifically, the timestamps of the open, high, low, and close prices within each bar. In this paper, we investigate the impact of incorporating this timing data into machine learning models for predicting volume-weighted average price (VWAP). Our experiments show that including these features consistently improves predictive performance across multiple ML architectures. We observe gains across several key metrics, including log-likelihood, mean squared error (MSE), $R^2$, conditional variance estimation, and directional accuracy.

Bergault, Philippe; Bieber, Sébastien; Guéant, Olivier; Zhang, Wenkai

doi: 10.48550/arxiv.2509.03964pmid: N/A

Abstract:In traditional financial markets, yield curves are widely available for countries (and, by extension, currencies), financial institutions, and large corporates. These curves are used to calibrate stochastic interest rate models, discount future cash flows, and price financial products. Yield curves, however, can be readily computed only because of the current size and structure of bond markets. In cryptocurrency markets, where fixed-rate lending and bonds are almost nonexistent as of early 2025, the yield curve associated with each currency must be estimated by other means. In this paper, we show how mathematical tools can be used to construct yield curves for cryptocurrencies by leveraging data from the highly developed markets for cryptocurrency derivatives.

Zhao, Yun; Zheng, Harry

doi: 10.48550/arxiv.2509.26535pmid: N/A

Abstract:We propose an approach to applying neural networks on linear parabolic variational inequalities. We use loss functions that directly incorporate the variational inequality on the whole domain to bypass the need to determine the stopping region in advance and prove the existence of neural networks whose losses converge to zero. We also prove the functional convergence in the Sobolev space. We then apply our approach to solving an optimal investment and stopping problem in finance. By leveraging duality, we convert the nonlinear HJB-type variational inequality of the primal problem into a linear variational inequality of the dual problem and prove the convergence of the primal value function from the dual neural network solution, an outcome made possible by our Sobolev norm analysis. We illustrate the versatility and accuracy of our method with numerical examples for both power and non-HARA utilities as well as high-dimensional American put option pricing. Our results underscore the potential of neural networks for solving variational inequalities in optimal stopping and control problems.