Applications of synthetic financial data in portfolio and risk modeling

TL;DR

This study evaluates TimeGAN and Variational Autoencoders for generating synthetic SP500 daily returns and demonstrates that TimeGAN closely replicates distributional shapes and temporal dynamics, enabling realistic downstream portfolio optimization and risk modeling. VAEs offer stable training but tend to smooth tail events, influencing risk estimation. Overall, synthetic data provide privacy-preserving, reproducible avenues for portfolio analysis, stress testing, and model development, with TimeGAN delivering the strongest empirical fidelity across fidelity, temporal coherence, and utility metrics.

Abstract

Synthetic financial data offers a practical way to address the privacy and accessibility challenges that limit research in quantitative finance. This paper examines the use of generative models, in particular TimeGAN and Variational Autoencoders (VAEs), for creating synthetic return series that support portfolio construction, trading analysis, and risk modeling. Using historical daily returns from the S and P 500 as a benchmark, we generate synthetic datasets under comparable market conditions and evaluate them using statistical similarity metrics, temporal structure tests, and downstream financial tasks. The study shows that TimeGAN produces synthetic data with distributional shapes, volatility patterns, and autocorrelation behaviour that are close to those observed in real returns. When applied to mean-variance portfolio optimization, the resulting synthetic datasets lead to portfolio weights, Sharpe ratios, and risk levels that remain close to those obtained from real data. The VAE provides more stable training but tends to smooth extreme market movements, which affects risk estimation. Finally, the analysis supports the use of synthetic datasets as substitutes for real financial data in portfolio analysis and risk simulation, particularly when models are able to capture temporal dynamics. Synthetic data therefore provides a privacy-preserving, cost-effective, and reproducible tool for financial experimentation and model development.

Figures (6)

• Figure 1: End-to-end pipeline for synthetic financial data generation and downstream evaluation.
• Figure 2: Comparison of portfolio weights between real and synthetic S&P 500 data.
• Figure 3: Comparison of volatility, Value-at-Risk (VaR$_{0.95}$), and Expected Shortfall (ES$_{0.95}$) between real S&P 500 returns and synthetic datasets.
• Figure 4: Comparison of the marginal distribution of real and synthetic log-returns.
• Figure 5: Autocorrelation and DTW comparisons for real and synthetic financial series.