Guosen: The final mindset of asset trend convergence.

date
08:50 08/02/2026
avatar
GMT Eight
In the past 20 years, global asset classes have experienced two periods of high correlation, the first being in mid-2013 and the second being in the first quarter of 2020 when the global outbreak of COVID-19 caused a rush to US dollar cash.
Guosen released a research report stating that in 2023, the global correlation of major asset classes once again reached a new high. The reasoning behind this lies in the end of the Kangbo recession, geopolitical risks, and under the AI narrative, the combination of "technology stocks + precious metals" trending stronger. In addition, with the convergence of the monetary cycles of China and the United States, diversifying risks becomes more difficult. Over the past 20 years, similar high correlations have only occurred twice: in 2013 due to taper tantrum fears triggering expectations of a liquidity turning point, and in 2020 due to the COVID-19 pandemic causing a USD liquidity crisis. The main points of Guosen are as follows: By measuring the correlations of similar assets priced in China and overseas over a period of time to evaluate the synergy of global major asset classes, the correlation has reached a new high in 2023. The logic behind the resonance of global major asset classes stems from the end of the Kangbo recession, the context of grand geopolitical events such as the Ukraine crisis and global trade frictions, the AI narrative brought by the wave of artificial intelligence, and the pursuit of safe-haven assets driven by the lack of confidence in the US dollar, forming a dual combination of "technology stocks + precious metals". Especially, the correlation between commodities and stocks has reached a new high, and the underlying monetary cycles behind bonds are evolving from divergence to synchronization. In the context of global asset consistency, it has become particularly difficult to diversify risks and reduce volatility across assets. There have been two instances of high global correlation of major asset classes over the past 20 years. The first occurred in mid-2013, and the second was in the first quarter of 2020 during the global COVID-19 outbreak, where cash was king. The high correlation of assets in mid-2013 was driven by fears of a global liquidity turning point triggered by expectations of a change in Fed policy, known as the taper tantrum. The sharp rise in the 10-year US Treasury yield within a short period of time led to a sharp reassessment of all long-dated assets, causing gold, US bonds, emerging market stocks, and bonds to simultaneously decline. The high correlation at this time was not due to fundamental resonance, but rather a systemic discount in anticipation of tightening liquidity, as the market transitioned from abundant liquidity to monetary normalization. The extreme risk aversion and USD liquidity crisis in the first quarter of 2020 resulted in the same direction of movement of all assets globally. During this phase, the characteristics were "cash is king", and the high correlation of assets was dominated by trading behavior of short selling everything to acquire USD liquidity. Following the outbreak of the COVID-19 pandemic, the market experienced a shift from concerns of fundamental recession to depletion of liquidity in the financial markets. In mid-March, the fear index VIX rose to a historical high of 85.47, with the US stock market halting trading four times within two weeks. The key turning point was when safe-haven assets failed, with gold falling by more than 10% between March 9 and 18, and even US bonds experiencing price declines due to institution selling for additional margin. At this time, the high correlation of global major asset classes was due to all assets being forced to act as liquidity realization tools, with fundamental logic temporarily giving way to micro capital structural deleveraging pressure, until the Fed implemented unlimited QE to break this positive correlation chain. Risk warning: Asset correlation phenomena are rare, and historical analysis does not represent the future.