Dependency Network-Based Portfolio Design with Forecasting and VaR Constraints

2025-07-26

AI Summary

This paper addresses the limitation of conventional portfolio optimization in capturing dynamic market interdependencies. Methodologically, it proposes a dynamic asset allocation framework integrating statistical network analysis and time-series forecasting: (i) a directed stock influence network is constructed using Vector Autoregression (VAR) and Forecast Error Variance Decomposition (FEVD); (ii) the Minimum Spanning Tree (MST) algorithm extracts systemic core structure from this network; (iii) returns are forecast via hybrid ARIMA–Neural Network AutoRegressive (NNAR) models, and risk-adjusted portfolio weights are optimized using Value-at-Risk (VaR). The key contribution lies in the novel coupling of FEVD-driven directed dependency networks with MST-based topological identification, enabling dynamic integration of network-structure awareness and risk-sensitive decision-making. Empirical evaluation over a one-year backtest shows that the NNAR-enhanced MST strategy achieves 63.74% cumulative return—substantially outperforming the buy-and-hold benchmark (18.00%)—demonstrating superior adaptability and alpha generation capability.

Technology Category

• Reasoning under Uncertainty: Stochastic Optimization
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Models for Web evolution
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

This study proposes a novel portfolio optimization framework that integrates statistical social network analysis with time series forecasting and risk management. Using daily stock data from the S&P 500 (2020-2024), we construct dependency networks via Vector Autoregression (VAR) and Forecast Error Variance Decomposition (FEVD), transforming influence relationships into a cost-based network. Specifically, FEVD breaks down the VAR's forecast error variance to quantify how much each stock's shocks contribute to another's uncertainty information we invert to form influence-based edge weights in our network. By applying the Minimum Spanning Tree (MST) algorithm, we extract the core inter-stock structure and identify central stocks through degree centrality. A dynamic portfolio is constructed using the top-ranked stocks, with capital allocated based on Value at Risk (VaR). To refine stock selection, we incorporate forecasts from ARIMA and Neural Network Autoregressive (NNAR) models. Trading simulations over a one-year period demonstrate that the MST-based strategies outperform a buy-and-hold benchmark, with the tuned NNAR-enhanced strategy achieving a 63.74% return versus 18.00% for the benchmark. Our results highlight the potential of combining network structures, predictive modeling, and risk metrics to improve adaptive financial decision-making.

Problem

• Integrates network analysis with forecasting for portfolio optimization
• Uses VAR and FEVD to quantify stock influence and risk
• Demonstrates MST-based strategies outperform traditional buy-and-hold

Innovation

• Dependency networks via VAR and FEVD analysis
• Minimum Spanning Tree extracts core stock structure
• ARIMA and NNAR forecasts refine stock selection