Risk-at-Risk: A Quantile-Based Indicator of Diversification Vulnerability

Investors have long relied on a simple principle to protect their wealth: do not put all your eggs in one basket. By holding a mix of different assets, such as stocks from various countries alongside gold or government bonds, they hope that if one part of their portfolio suffers, another will hold steady or even rise. This strategy, known as diversification, works best when the assets move independently of one another. However, financial markets are not static; the relationships between assets change, especially during times of stress. When a crisis hits, assets that usually behave differently can suddenly start moving in lockstep, causing the safety net to fail just when it is needed most. Understanding when and why these protective assets lose their ability to shield a portfolio is a central challenge for anyone managing money, from individual savers to large institutions.

A new study introduces a fresh way to monitor this vulnerability, focusing not on how individual assets move together, but on how the internal risk within two distinct groups of assets evolves in relation to each other. The researchers distinguish between "risky assets," which are primarily stock market indices from around the world, and "hedge assets," which include traditional safe havens like gold, the US dollar, and government bonds. Under normal conditions, these groups play different roles. But the study asks a specific question: when the stocks within the risky group begin to cluster in their worst moments, does the risk within the safe-haven group also begin to cluster in a similar way? If both groups become internally chaotic at the same time, the diversification strategy breaks down because the two groups are no longer providing distinct sources of protection.

To answer this, the authors developed a dynamic indicator they call "Risk-at-Risk." Instead of looking at the average behavior of markets, which often hides extreme events, they focused specifically on the worst days. They first measured how tightly connected the worst days were within the group of stocks and separately within the group of safe assets. Then, they tracked how these two internal states of stress moved together over time. If the stocks are getting more tightly linked in their downturns, and at the same time the safe assets are also getting more tightly linked in their own downturns, the indicator rises. A high value signals that the two groups are aligning in their vulnerability, meaning a portfolio holding both might offer less protection than expected.

The researchers analyzed daily data spanning from 1997 through July 2026, covering a wide range of global markets. They found that this relationship is not constant; it changes significantly over time. The indicator showed strong positive alignment during the global financial crisis, the early 2010s, and again around 2017 and 2018. These were periods where the internal risk structures of both stocks and safe assets became closely synchronized. Interestingly, the study found that this alignment does not always happen when markets are simply volatile. For instance, during the peak of the pandemic crisis, while the risk within the stock market surged, the relationship between the two groups did not strengthen in the same way. This suggests that a rise in stock market fear does not automatically mean that safe assets are losing their protective power; the two phenomena are distinct.

A crucial finding of the research is that looking at the average behavior of markets tells a different story than looking at the worst days. The researchers compared their new indicator with traditional measures that look at the middle of the market distribution. They found that the relationship between the two groups during bad times often moves in the opposite direction of the relationship during normal times. In the years following the mid-2010s, for example, the average relationship between the groups became negative, suggesting that on typical days, the groups were moving apart. Yet, during the worst days, the relationship remained positive, indicating that the groups were still aligning when it mattered most. This divergence highlights that relying on average correlations can be misleading when assessing true diversification vulnerability.

The study also revealed important differences between advanced economies and emerging markets. While both groups showed similar turning points during major crises, the emerging markets exhibited much larger swings in the indicator. The relationship in emerging markets tended to spike more sharply and recover faster after crises, reflecting the higher volatility and quicker shifts in investor confidence typical of those regions. This suggests that investors in emerging markets face a more volatile diversification landscape, where the safety of their hedges can change more rapidly than in more mature economies.

By focusing on the alignment of internal risk structures rather than just direct connections between individual assets, this new indicator offers a more nuanced view of portfolio safety. It suggests that the failure of diversification is often a result of two separate groups of assets becoming internally chaotic at the same time, rather than just one group crashing into the other. The research provides a tool for monitoring these subtle shifts, helping investors understand that the safety of a hedge is not a fixed property but a dynamic condition that can fade when the internal risk of the entire market system becomes too concentrated.