COVID-19 Triggers Unique Market State in S&P 500
Dr. Manan Vyas's team found that correlations between S&P 500 stocks entered a previously unseen market state a few months into the COVID-19 pandemic, which persisted for over a year.
Read the original paper
Researchers from the Instituto de Ciencias Físicas, Universidad Nacional Autónoma de México, and the Centro Internacional de Ciencias AC have identified a unique market state in the S&P 500 index correlated with the COVID-19 pandemic. The study, led by Dr. Manan Vyas and conducted with her team including Mijaíl Martínez-Ramos, Parisa Majari, and Professor Thomas Seligman, analyzed market behavior over a period of more than 15 years and discovered a novel state during the pandemic period. Their findings are published in the peer-reviewed journal PLOS ONE.
Analyzing the correlation of returns, the researchers observed a market state that was unprecedented in the historical data spanning over 15 years. This “COVID state” emerged following the initial financial panic of the pandemic and exhibited distinct characteristics not previously seen in the market. The study utilized the Pearson correlation matrix and relative correlations with respect to the S&P 500 index to identify this anomaly.
“The COVID state does not appear immediately after the pandemic’s onset but a few months later. This period corresponds to a crash well identified with panic sell-off,” said Dr. Vyas. This lag between the onset of the pandemic and the appearance of the COVID state is attributed to the initial market panic which masked the subtler correlations that emerged later.
The team’s analysis revealed that the COVID state was characterized by a low average correlation and a distinct clustering in the correlation matrix space. This state persisted for over a year, indicating a significant shift in market dynamics during the pandemic. The researchers employed k-means clustering to categorize the market states and confirmed the robustness of the COVID state across different numbers of clusters.
Dr. Vyas explained, “We found that a previously non-existent market state appears in a time frame closely related to the COVID pandemic. This suggests that the economic consequences and the mindset of the population during COVID had a profound impact on market behavior.”
Further analysis showed that the transition matrices during the COVID state were nearly tri-diagonal, reinforcing the uniqueness of this market condition. The equilibrium distributions indicated a distinct pattern of market behavior that was not observed in other periods.
The study’s findings have significant implications for understanding market dynamics during extreme events. The researchers emphasized the importance of considering relative correlations, which provided deeper insights into the market states by filtering out the dominant effects of the average correlation.
In conclusion, the appearance of the COVID state highlights the need for advanced analytical tools to detect and understand new market conditions. As Dr. Vyas noted, “This example will help us in our search for relevant parameters in the stock market beyond the highest eigenvalue of the Pearson correlation matrix.”
About the researchers
Journal reference
Martínez-Ramos, M. M., Vyas, M., Majari, P., & Seligman, T. H. “COVID anomaly in the correlation analysis of S&P 500 market states.” PLOS ONE (2024): 19(4): e0301238. https://doi.org/10.1371/journal.pone.0301238





