Bloomberg Professional Services
- Correlations reveal the true market regime: Unlike volatility or spreads, cross-asset correlations capture how risks are interconnected and whether markets are driven by common forces, providing a more structural view of market behavior.
- Similar crises can have different underlying structures: While episodes like 2008 and 2020 both exhibit risk-off dynamics, their correlation patterns differ due to distinct drivers, underscoring the importance of looking beyond surface-level similarities.
- The current regime reflects a structural shift driven by inflation: Markets are moving away from past crisis patterns toward a new environment where inflation and policy expectations drive stronger co-movement across asset classes, reshaping traditional diversification.
This article was written by Antonios Lazanas, Head of Quantitative Investment Research, and Changxiu “Sue” Li, Head of Fixed Income Portfolio Analytics at Bloomberg.
A structural view of market regimes
Analysts tend to describe the state of the market using levels – spreads, volatility, or valuations. These measures are intuitive, but they lack the power to fully characterize the true nature of a particular market environment. Periods that appear similar on the surface may behave very differently underneath.
PRODUCT MENTIONS
A more robust way to define a market regime is to examine how assets move relative to one another. Asset correlation provides a direct measure of these relationships, capturing the extent to which different parts of the market are driven by common forces. In this sense, correlation reflects the structure of risk rather than its magnitude.
This distinction is critical. Volatility describes how much markets move. Correlation describes whether they move together. Two environments with similar volatility can exhibit very different levels of systemic risk depending on how tightly assets are linked.
Methodology: Measuring regime similarity
To formalize this idea, we use Bloomberg’s MAC3 monthly global risk model and a set of 16 factors covering equities, inflation, commodities, foreign exchange, credit, and rates. These factors are designed to represent the key risk transmission channels of global markets.
The MAC3 monthly model is calibrated on a rolling basis and typically reflects an effective estimation half-life of approximately 10 to 15 months. As a result, the model is particularly well suited to capturing current and recent regime dynamics, rather than long-run structural averages.
We use the month-end factor correlation matrices from 2001 to 2026. We then measure the distance between correlation matrices using a Euclidean (Frobenius) metric and project the results into a three-dimensional space via Multi-Dimensional Scaling (MDS). Each axis represents a latent dimension summarizing similarities in the cross-asset correlation structure. In this representation, proximity reflects similarity in cross-asset structure.
Figure 1: Market regime evolution 2001-2026
Mapping the evolution of market structure
The MDS representation reveals a clear separation of market regimes over time. The pre-2008 period, particularly from 2005 to 2007, forms a relatively tight cluster characterized by stable and loosely connected cross-asset relationships. Inflation played a limited role, and interactions between rates, credit, and other macro factors were relatively muted.
Crisis periods such as 2008 and 2020 appear as distinct outliers, reflecting a sharp compression in correlations as markets transition into synchronized risk-off behavior. Diversification breaks down, and asset classes become increasingly driven by a common systemic factor.
At the same time, the representation highlights a more nuanced relationship between these two crises. The 2008 and 2020 observations align closely along two dimensions of the embedding (MDS1 and MDS2), reflecting a shared flight-to-quality dynamic. In both episodes, risk assets sold off while safe-haven assets rallied, producing a similar cross-asset correlation signature.
However, the two regimes do not fully overlap. The separation across other dimensions reflects important structural differences, most notably the role of mortgage and securitized credit factors during the Global Financial Crisis. These factors were central to the 2008 dislocation but largely absent in the 2020 shock, resulting in a distinct overall correlation structure despite similar directional moves.
Another notable feature of the map is the position of the 2022 inflation episode. While not traditionally classified as a “stress” period in the same sense as 2008 or 2020, it appears as a distinct extreme in the correlation space.
This reflects a fundamentally different type of regime. Rather than a broad risk-off event, 2022 was characterized by a reconfiguration of relationships between rates and inflation. Assets that historically provided diversification — particularly fixed income — became positively correlated with equities as inflation shocks drove both markets simultaneously.
The result is a correlation structure that is highly unusual relative to both crisis and pre-crisis environments. This underscores that not all regime shifts are driven by stress; some are driven by structural changes in macro dynamics.
Where are we today?
The AI-driven surge in market optimism has fueled a relentless rally in public equities and a boom in private markets over the past several years. As valuations rise, investors are beginning to question whether markets are once again slipping into what American economist Robert J. Shiller famously called “irrational exuberance.” Some analysts draw parallels to the dot-com boom and bust of 1999–2001. Others point to tight credit spreads and rapid growth in private credit as warning signs of a potential crisis reminiscent of 2008. Still others highlight the risk of stagflation—a scenario that has gained renewed attention following the Iran war and the associated oil supply shock.
To better understand the current environment, consider the position of recent1 correlation matrices in the three-dimensional space shown in Figure 1. After inflation fears began to ease in 2023, markets gradually moved toward the strong-growth cluster observed during 2009–2019, although elevated inflation risk kept this regime distinct from other growth periods. This trajectory shifted sharply in 2026, when the Iran war and ensuing oil shock altered market dynamics.
Importantly, this shift does not resemble a move toward the 2008-style credit crisis. Instead, it points toward a previously sparse region between the dot-com period of 2001 and the inflation-driven environment of 2022. This suggests a structural change in markets. Inflation risk has become a central driver of cross-asset behavior: rates and credit now move more closely together, and asset prices are increasingly shaped by policy expectations. These characteristics stand in contrast to the pre-2008 environment, which was defined by low inflation and relatively weak cross-asset linkages.
Leveraging the accuracy and responsiveness of our MAC3 global risk model covariance estimation methodology, we will continue to monitor these dynamics closely and track how this regime evolves.
Conclusion
Market regimes cannot be fully understood through traditional metrics alone. By focusing on cross-asset correlations, we gain a clearer view of the underlying structure of risk and how it evolves over time. This perspective reveals that not all periods of stress—or stability—are alike, and that structural shifts can emerge even outside of crisis environments.
Today’s market appears to be entering a distinct regime shaped by inflation dynamics and policy expectations, with implications for diversification and asset behavior. As these relationships continue to evolve, maintaining a structural, data-driven lens will be essential to interpreting market signals and navigating uncertainty.
1 The last correlation matrix in Figure 1 is as of March 13, 2026