Research Review | 5 June 2026 | Risk Management

Key Takeaways

• Put‑call disparity offers model‑free bubble bounds without parametric assumptions
• Correlation spikes in 71% of crises, weakening traditional diversification
• Trachsler Resilience Score cuts drawdowns in 83% of stress events
• Market‑state similarity strategy outperforms equal‑weighted industry benchmark
• Regime‑aware RL allocation improves risk‑adjusted returns across equities, Treasuries, gold

Pulse Analysis

The new model‑free bubble measurement introduced by Jarrow and Kwok sidesteps the need for restrictive parametric assumptions, relying only on no‑free‑lunch‑with‑vanishing‑risk and admissible trading strategies. By exploiting observable put‑call disparities and the lowest out‑of‑the‑money call price, the framework delivers credible lower and upper bounds on bubble size. Empirical tests on S&P 500 options trace a sustained COVID‑era bubble and correctly flag the pre‑crash exuberance of the 2000 dot‑com and 2008 financial crises, offering regulators and investors a practical early‑warning signal.

Diversification under stress is another focal point, as Trachsler’s extensive analysis of sector ETFs and international indices reveals a systematic correlation breakdown during systemic shocks. Mean pairwise correlations rise in 71.2% of identified crises, eroding the risk‑reduction benefits of naive diversification. The newly proposed Trachsler Resilience Score selects sectors that remain loosely coupled, delivering lower maximum drawdowns in 83% of out‑of‑sample stress events. Complementary research on sector ETFs shows that divergence regimes, such as the AI‑boom of 2022‑2026, can be identified and exploited through targeted momentum sleeves, underscoring the need for dynamic, correlation‑aware portfolio construction.

Beyond bubbles and diversification, the collection spotlights sentiment‑driven mispricing, industry‑rotation tactics, and regime‑aware allocation. Ryu et al. demonstrate that positive and negative consumer confidence differentially affect returns and volatility across developed and emerging markets. Zakamulin’s market‑state similarity approach extracts tail‑risk information from lagged market returns, delivering higher Sharpe ratios and reduced drawdowns versus passive benchmarks. Finally, Verma et al. blend hidden Markov models with reinforcement learning to allocate across SPY, TLT and GLD, dynamically shifting exposure as regimes transition. Together, these studies equip asset managers with quantitative, data‑driven tools—reinforced by practical resources like the new R‑based portfolio analytics guide—to navigate an increasingly complex risk landscape.