[Paper Review] Dynamic Asset Allocation with Asset-Specific Regime Forecasts

Yizhan Shu, Chenyu Yu+1 more|arXiv (Cornell University)|Jun 13, 2024

Banking stability, regulation, efficiency5 citations

TL;DR

The paper introduces an asset-specific regime forecasting framework that uses unsupervised statistical jump models to label regimes for each asset, followed by supervised forecasting and Markowitz optimization to improve multi-asset allocation.

ABSTRACT

This article introduces a novel hybrid regime identification-forecasting framework designed to enhance multi-asset portfolio construction by integrating asset-specific regime forecasts. Unlike traditional approaches that focus on broad economic regimes affecting the entire asset universe, our framework leverages both unsupervised and supervised learning to generate tailored regime forecasts for individual assets. Initially, we use the statistical jump model, a robust unsupervised regime identification model, to derive regime labels for historical periods, classifying them into bullish or bearish states based on features extracted from an asset return series. Following this, a supervised gradient-boosted decision tree classifier is trained to predict these regimes using a combination of asset-specific return features and cross-asset macro-features. We apply this framework individually to each asset in our universe. Subsequently, return and risk forecasts which incorporate these regime predictions are input into Markowitz mean-variance optimization to determine optimal asset allocation weights. We demonstrate the efficacy of our approach through an empirical study on a multi-asset portfolio comprising twelve risky assets, including global equity, bond, real estate, and commodity indexes spanning from 1991 to 2023. The results consistently show outperformance across various portfolio models, including minimum-variance, mean-variance, and naive-diversified portfolios, highlighting the advantages of integrating asset-specific regime forecasts into dynamic asset allocation.

Motivation & Objective

• Motivate improvements in first-stage forecasting by leveraging asset-specific market regimes rather than broad economic regimes.
• Develop a two-step regime framework (identification then forecasting) to produce live-ready regime forecasts.
• Integrate regime forecasts into standard portfolio optimization models (minimum-variance, mean-variance, and equally-weighted).
• Demonstrate robustness and outperformance across a diverse asset universe including equities, fixed income, real estate, and commodities.

Proposed method

• Apply an unsupervised statistical jump model (JM) to classify historical periods for each asset into bullish or bearish regimes.
• Extract eight asset-return features with exponential smoothing; compute a two-regime JM with a jump penalty lambda to control persistence and signal-to-noise ratio.
• Shift identified regime labels forward by one day to serve as targets for a supervised classifier.
• Train a gradient-boosted decision tree (XGBoost) to forecast the next-period regime using expanded feature sets.
• Incorporate regime forecasts into Markowitz mean-variance optimization to determine dynamic asset allocation weights across the asset universe.
• Evaluate performance via out-of-sample testing including transaction costs (one-way 5 basis points) across three portfolio models (MinVar, MV, EW).

Experimental results

Research questions

• RQ1 Can asset-specific regime forecasts improve the predictive signal beyond broad economic regimes for multi-asset allocation?
• RQ2 Does combining unsupervised regime identification with supervised forecasting yield more persistent and actionable regime signals for each asset?
• RQ3 Do regime-informed allocations outperform traditional portfolio models under realistic trading costs across a diversified asset universe?
• RQ4 How does the optimal jump penalty affect regime signal quality and subsequent forecast accuracy across assets?

Key findings

• Regime forecasts generated asset-by-asset capture distinct market dynamics and lead to improved forecasting signals for subsequent optimization.
• The framework yields outperformance across minimum-variance, mean-variance, and equally-weighted portfolios in an empirical study of twelve assets from 1991–2023.
• Outperformance persists under realistic transaction costs and diverse asset classes, including equities, bonds, real estate, and commodities.
• Asset-specific regime forecasts provide robustness by avoiding over-reliance on a single market regime narrative for all assets.