SONAR|HES-SO

SONAR|HES-SO

SONAR|HES-SO regroupe les travaux de bachelor et master diffusables de plusieurs écoles de la HES-SO. Consultez cette page pour le détails.

En cas de question, merci de contacter les bibliothécaires de la HES-SO : bibliotheques(at)hes-so.ch

Back to SONAR
Back
ILAZI-ARBER-FINAL-JUL.pdf
Bachelor thesis

Navigating peaks and troughs : can we leverage volatility metrics in passive dynamic asset allocation ?

  • Genève : Haute école de gestion de Genève

88 p.

Bachelor of Science HES-SO en Economie d’entreprise: Haute école de gestion de Genève, 2024

English This thesis explores Volatility-Based Dynamic Asset Allocation (DAA) strategies with the goal of analyzing various volatility metrics, evaluating their predictive power, and developing models to enhance portfolio performance and risk management.
Part I analyzes current S&P500 market dynamics. This section provides an analysis of various volatility metrics and of the CBOE VIX Index (VIX). Part II conducts an empirical study on the predictability of the VIX, by comparing the VIX with various calculated volatilities (historical volatilities and GARCH models). The Variance Premium (VP), which predicts future volatility, is calculated by subtracting rolling historical volatility from implied volatility. Part III develops volatility-based DAA models using R. An ETF portfolio is constructed with portfolio re-allocation rules defined for different volatility regimes and metrics. Each end-strategy is composed of one allocation set and one volatility metric.
• Sets of allocations: MONTBLANC (initial allocations) and ZERMATT (designed
to improve risk metrics).
• Strategies: BASIC (re-allocation based on VIX levels), PEAK (re-allocation based
on 252-day VP levels) and EDGE (re-allocation based on 21-day VP levels).
Key Findings:
VIX Predictability: The results show that the VIX’s predictive power is limited due to its strong correlation with short-term historical volatility. Comparing the VIX to GARCH models proved that complex models offered little improvement. When the VP is calculated using short-term volatility – given that the VIX closely tracks it – it evaluates the predictive value in the small difference between the VIX and the short-term historical volatility, essentially becoming the difference between short and long-term historical volatility.
Backtest results: PEAK provides the best Sharpe Ratios, often outperforming all benchmarks. BASIC offers a balanced approach but with moderate results. EDGE showed poor returns with high volatility. Considering end-strategies, MONTBLANC PEAK shows the highest results, outperforming all benchmarks out-sample. We also optimized allocation sets and volatility thresholds, both showing robust out-sample results, which is uncommon as optimizations often underperform on future data.
Conclusion: This study emphasizes the importance of integrating various volatility metrics to develop more resilient portfolios. Future research could explore more sophisticated models and include behavioral finance aspects and active decision-making to build more effective and adaptative portfolio management strategies.
Language
  • English
Classification
Economics
Notes
  • Haute école de gestion Genève
  • Economie d’entreprise
  • hesso:hegge
Persistent URL
https://sonar.ch/hesso/documents/331314
Statistics
Document views: 129 File downloads:
  • ILAZI-ARBER-FINAL-JUL.pdf: 653