Market correlations don’t stay constant during financial crises. Assets that usually move independently often start moving together, reducing the effectiveness of diversification. For example, during the 2008 crisis, equity correlations jumped from 0.35 to over 0.80. Similarly, in the 2020 COVID-19 crash, correlations spiked to 0.75 in just a few months.

This happens because fear and liquidity pressures drive investors to sell across the board, causing asset prices to sync up. Traditional portfolio models, which assume stable correlations, often underestimate risks by 40–60% during such times.

To manage this, traders need to:

• Use dynamic models that account for crisis-level correlations. Many traders learn these techniques through a systematic traders program designed for volatile regimes.
• Stress-test portfolios with higher correlation scenarios.
• Include hedging strategies like cash or long-dated put options.

Systematic trading strategies don’t work when they remain static during market volatility. Recognising these shifts and preparing for them is critical for protecting your portfolio during crises.

How Do Correlations Between Assets Change During Major Financial Crises

sbb-itb-466c9b0

Past Financial Crises and Correlation Changes

The 1998 Global Financial Crisis

The 1998 financial crisis, which began with the Asian Financial Crisis in July 1997 and was later exacerbated by Russia’s financial collapse and Brazil’s currency challenges, brought a sharp rise in global market comovement. Fewer dominant factors drove market movements, amplifying cross-market feedback. A study of 13 Asian and non-Asian markets found that cross-market interactions increased seven-fold, meaning volatility in one market could more easily ripple through others. This surge in interdependence was largely driven by temporary shocks rather than any lasting structural shifts.

These developments laid the groundwork for future crises, where similar patterns of heightened market interconnection would reappear.

The 2008 Global Financial Crisis

The 2008 crisis saw global markets moving almost in unison, as high volatility led to a significant rise in correlations. Researchers Leonidas Sandoval and Italo De Paula Franca explained this phenomenon:

High volatility of markets is directly linked with strong correlations between them. This means that markets tend to behave as one during great crashes.

The U.S. stock market played a central role, with its returns influencing and predicting movements in European and Australasian markets. During this period, emerging markets, which often move independently during stable times, became tightly linked with developed markets, showcasing the global contagion effect. Statistical analysis revealed that Asian markets (excluding Japan) formed a distinct cluster, while other global markets grouped into a separate correlated cluster. Gulser Meric and colleagues noted:

Correlation between global stock markets has increased and the benefit of global portfolio diversification has decreased since the 2008 stock market crash.

These findings highlighted the diminishing advantages of diversification during periods of crisis.

The COVID-19 Pandemic

The COVID-19 pandemic introduced a major shift in market correlations, breaking patterns observed in previous crises. For 12 years after the 2008 crisis, the correlation of U.S. asset returns remained relatively stable at around –0.30. However, this stability was disrupted by the pandemic. Initially, developed economies showed signs of isolation, but this was soon followed by a steady rise in correlation measures across nearly all asset pairs. Researchers also found notable increases in correlations between China’s market and commodities like oil and gas, as well as between China and Italy.

Unlike earlier crises, where correlation spikes were temporary, the post-COVID environment has seen sustained increases in multi-asset correlations. This shift has challenged long-standing diversification strategies. LSEG researchers warned:

Evidence since 2021-22 confirms that relying on historical correlations to predict future correlations may give poor results.

Adding to this, the correlation between U.S. stock and government bond returns surged in 2022–2023 as inflation climbed and central banks raised interest rates. This development undermined the traditional role of bonds as a hedge against equities, calling into question the effectiveness of the 60/40 portfolio model that had been a cornerstone of investment strategy for decades.

What Drives Correlation Changes During Crises

How Investor Sentiment Affects Correlations

During crises, fear often overrides logical decision-making. Investors tend to sell off diversified assets indiscriminately, regardless of whether those assets are fundamentally connected. This widespread selling leads to a phenomenon where asset prices become unusually correlated. Shawn Mankad, an Assistant Professor at Cornell University, described the 2008 financial crisis as a time when:

stock prices became super correlated because everyone was selling their stocks at the same time.

He further explained that:

the correlation network was a big ball of connections.

This behaviour sheds light on why emerging markets, which usually operate independently during stable times, suddenly sync up with developed markets during periods of turmoil. Research consistently shows that correlation levels spike during major crises, such as the Global Financial Crisis and the European Sovereign Debt Crisis, providing clear evidence of contagion. These shifts eliminate the usual benefits of diversification and worsen liquidity pressures in already stressed markets.

Liquidity Crises and Their Effects

When liquidity dries up, institutional investors face tough choices. Margin calls or mounting redemption requests force fund managers to sell assets quickly – not necessarily the ones they would prefer to sell. This leads to what experts call a “wealth effect”, where losses in one market push institutions to liquidate holdings in unrelated markets to raise cash.

This explains why asset prices can move together, not because of any fundamental link, but because they’re all being sold at the same time. For example, during the 2008 crisis, European banks stopped lending to one another, while their stock prices became highly correlated. This demonstrates how physical market disruptions, like the interbank lending freeze, can trigger sharp increases in correlation across financial markets. Liquidity-driven sell-offs also make markets more vulnerable to sudden shifts caused by breaking news.

Time-Varying Correlations and News Events

Correlations are not fixed – they change in response to news, economic cycles, and systemic risks. Researchers using advanced models have observed significant shifts in conditional correlations during events like the Global Financial Crisis and the European Sovereign Debt Crisis. The COVID-19 pandemic offered another example: between May 2015 and August 2022, Dynamic Conditional Correlations between China and oil, as well as Italy and gas, increased notably.

Macroeconomic factors play a major role. Inflation, for instance, is a key driver of higher correlations. When inflation and interest rates rise, the traditional negative correlation between stocks and bonds often disappears. The LSEG research team pointed out that:

correlation breakdowns may reflect time-varying volatility of financial markets rather than a change in the relationships between asset returns.

This means that higher market volatility alone can cause correlations to appear stronger, even if the underlying relationships between assets remain unchanged.

Markets also react unevenly to news. Downside risks, such as spillovers between global commodities and emerging market stocks, tend to have a much stronger contagion effect than positive news does. In downturns, panic amplifies these spillovers, making crises far more unpredictable. These dynamic shifts in correlations highlight the importance of staying vigilant and adjusting strategies during periods of market stress. Understanding these patterns is crucial for effectively navigating turbulent times.

How This Affects Systematic Trading

Systematic traders need to rethink how they approach diversification and risk management, especially in light of shifting market correlations during crises.

Adjusting Diversification During Crises

Diversification often falls apart during market crises. Under normal conditions, pairwise equity correlations usually stay around 0.30. But during systemic crises, these correlations can spike to 0.70 or higher. When this happens, the protective effect of diversification is significantly reduced.

Static models that assume constant correlations can leave portfolios vulnerable. These models may underestimate tail risk by as much as 40% to 60% compared to those that account for regime changes. For instance, during the October–November 2008 crash, correlations in developed markets soared from 0.35 to over 0.80.

To address this, systematic traders need to adopt dynamic models that account for different market regimes – distinguishing between calm and crisis periods. Crisis periods, once triggered, tend to last between 6 and 18 months and often begin with abrupt changes rather than gradual shifts. By stress-testing portfolios with crisis-level correlations of 0.75–0.80, traders can better evaluate whether their portfolios can withstand severe drawdowns.

Given the limitations of diversification during crises, risk management strategies must also be updated to safeguard portfolios.

Updating Risk Management Approaches

Risk management frameworks should be designed to handle the surge in correlations during turbulent markets. Recalculating Value at Risk (VaR) with correlations forced to 0.75–0.80 across equity positions can provide a clearer picture of potential risks. If these recalculations show a doubling of expected drawdowns, traders should immediately adjust their position sizes.

When diversification fails, structural hedges become essential. Allocating 2%–5% of a portfolio to long-dated put options or trend-following managed futures can offer protection that doesn’t rely on stable correlations. For example, during the 2020 COVID-19 crash, Nintai Investments LLC kept 15%–30% of partner portfolios in cash, which acted as a stabilising force when most asset classes moved in unison. Cash remains the only asset that guarantees both liquidity and stability when correlations approach 1.0.

Monitoring volatility indicators, like the VIX, can also provide early warnings of market stress. Traders might consider reducing equity exposure when volatility exceeds one standard deviation above its 12-month moving average. More advanced methods, such as spectral analysis of correlation matrices, can help track sharp increases in the leading eigenvalue, signalling emerging crises.

These adjustments in risk management underscore the importance of continuous learning, particularly in crisis management techniques.

Learning Resources from Collin Seow Trading Academy

The Collin Seow Trading Academy offers resources tailored to help traders manage market stress. Their Systematic Trader Programme focuses on building adaptive frameworks that respond to changing market conditions, moving away from static assumptions. The curriculum includes strategies for managing risk when traditional diversification falters and implementing alternative protective measures.

For those seeking to build foundational skills, the academy’s free Market Timing 101 E-Course teaches how to identify regime changes – an essential skill for anticipating correlation spikes. Additionally, the Systematic Trading Profits webclass introduces a three-phase system that incorporates dynamic allocation strategies. These tools equip traders to handle periods when asset prices move in unison, helping them safeguard their capital during critical market events.

Conclusion

Market correlations can change drastically during crises, pushing traders to rethink their approach to diversification and risk management.

During systemic events, correlations often spike to 0.70 or higher, eroding the benefits of diversification just when they’re needed most. Static models, which fail to account for these shifts, can underestimate tail risks by 40–60%, leaving traders exposed to major losses.

Historical examples like the 2008 Global Financial Crisis (correlations exceeding 0.80) and the 2020 COVID-19 pandemic (correlations near 0.75 within 2–3 months) demonstrate how static strategies fall short during turbulent periods. To navigate such conditions, traders need dynamic frameworks that adapt to both calm and crisis regimes. This includes stress-testing portfolios against crisis-level correlations and incorporating tail-risk hedging strategies.

These insights underscore the importance of adopting flexible strategies. For systematic traders, understanding the behaviour of correlations during crises is critical. Moving away from static models towards adaptive approaches demands a deeper level of education. Collin Seow Trading Academy addresses this need through its Systematic Trader Programme, which focuses on regime-switching models and dynamic risk management. Additionally, their free Market Timing 101 E-Course equips traders with the tools to anticipate market shifts and safeguard their portfolios effectively.

FAQs

How can I tell when correlations are starting to spike?

Monitoring sharp rises in metrics such as rolling correlation coefficients or eigenvalues of correlation matrices can help you spot spikes in correlations. These sudden changes often align with periods of market stress or increased volatility, which might indicate regime shifts or contagion effects. Paying close attention to these signals during turbulent market conditions can help you identify shifts early on.

Which assets still diversify well when markets move together?

Assets in less-integrated markets, such as some emerging and frontier markets in Africa and Asia, tend to show lower correlations with global benchmarks during periods of crisis. This means they often move independently of major markets, offering potential diversification benefits when global markets are closely aligned.

How do I stress-test my portfolio for crisis-level correlations?

To evaluate how your portfolio might perform under pressure, simulate scenarios where correlations between asset classes spike, often hitting levels of 0.70–0.80 during market crises. Adjust your correlation assumptions to reflect these conditions. It’s also worth exploring small allocations to tail-risk hedges, such as long-dated put options or managed futures, which can offer some protection in extreme market downturns.

Dynamic models, like regime-switching frameworks or rolling correlations, can be helpful tools. These models allow you to track changes over time, spot contagion risks early, and fine-tune your risk management strategies when markets become highly volatile.