In Part I of this series, we broke down in detail data about whether Chinese state owned enterprise banks were hoarding US dollars from the large ongoing trade surplus being run by China. According to Chinese banks themselves from their annual and quarterly reports, there is no evidence Chinese banks are rapidly increasing in either relative or absolute terms their holdings of USD or non-RMB assets in any form much less in liquid unencumbered forms.
Given the relationship between the Chinese state and the state owned nature of the banks involved for foreign currency transactions, does China believe that Chinese banks are hoarding US dollars? Put another way, while Chinese banks say they are not materially increasing their US dollar or non-RMB assets, does China have data that indicates a material difference from Chinese bank reports?
Due to the closed nature of the foreign exchange market in China, there is a government agency called the State Administration of Foreign Exchange that sets rules for foreign exchange transactions and collects data on foreign exchange transactions that do take place.
SAFE maintains a dataset called the External Assets and Liabilities in China’s Banking Sector with rather detailed data covering the data we are interested specifically the holdings of US dollars and non-RMB currencies. It provides detail on the major line items of interest like holdings of deposits, bonds, and break out by currency and location of holding.
According to SAFE between 2022 and 2025, the same time frame we used in Part I, non-RMB assets held by the Chinese banking sector within and outside China combined increased by a grand total of $169 billion USD. Chinese banking assets of US dollar assets specifically grew by a paltry $32 billion. To put that in perspective, in a time span when the Chinese current account surplus total $1.8 trillion USD, the China State Administration of Foreign Exchange Reserves records Chinese USD increasing by $32 billion or 1.7% of the aggregate current account surplus.
Beyond the snails pace of USD asset growth being recorded in Chinese banks, they are a few notable points. First, net assets grew relatively robustly because while gross asset growth was weak in non-RMB assets, net assets grew because non-RMB liabilities fell by $52 billion. This pushed net asset growth up by $220 billion USD.
Second, Chinese banks appear to be responding to interest rate differentials. While USD liabilities are down $190 billion Japanese Yen liabilities were the only currency which say increased liabilities though still minimal at $15 billion. This appears to be a pivot shifting from higher yield debt to lower yield liabilities.
Third, the change in net USD assets, overwhelmingly due to the drop in USD liabilities, is actually responsible for 100% of the change in total Chinese bank non-RMB net asset growth. This change in the net growth of USD asset in state owned banks of $222 billion matters because the big four state banks of China declared in their annual reported gross asset growth of $202 billion. Now for multiple reasons these are not the same number, for instance SAFE includes all Chinese banks, however it is notable that the numbers are at least coming close to one another.
However, there are other places within SAFE data we can find evidence about whether Chinese banks are boosting their holdings of USD assets. SAFE maintains a dataset on China’s international investment position. It must be noted and emphasized that this data goes well beyond the walls of Chinese state banks but we should be able to discern patterns about the behavior of the Chinese state.
Between 2022 and 2025, Chinese held deposits increased from $522 billion USD to $573 billion USD. In other words, during a time when Chinese companies were enjoying a $1.8 trillion USD surplus their deposits outside of China increased by only $51 billion USD or a total of 9.6% in four years. Nor did we witness growth in Chinese lending. Loans made by Chinese residents or companies increased from $850 billion USD in 2022 to $872 billion USD in 2025 growing a total of $22 billion in four years for a change of 2.6%.
There are other line items which Chinese banks could theoretically be in under SAFE international investment position, and other data we will be analyzing in the future, but given banks the world over and specifically the banks in question hold the large majority of their assets and liabilities as either deposits or loans, I will avoid going too far afield at the moment. We will however in the future revisit other channels where the Chinese surplus might be going.
Despite the argument that Chinese banks are acting as a defacto branch of the Chinese state and the People’s Bank of China becoming an unofficial repository for surplus US dollars, neither the Chinese banks themselves or Chinese regulators have data to support this argument. It must be emphasized that the differences here are not marginal differences that can be attributed to accounting or timing differences but rather many orders of magnitude.
As China was enjoying an official current account surplus of $1.8 trillion USD, officially China’s overseas loans grew by $22 billion with deposits growing by $51 billion. These are not marginal rounding errors but fundamental and foundational problems to believe Chinese banks act as a primary or even second repository for non-RMB assets.
The Chinese current account surplus exists. We can verify that data from multiple directions so the question is not whether the money from the current account surplus exists. The question is where did the Chinese current account surplus go? Despite arguments that it went to Chinese banks, there is simply no evidence that the Chinese current account surplus flowed in material amounts to Chinese state owned banks. Chinese banks do not record and nor do Chinese regulators.
A couple of final points. First, in the next few pieces we will focus on what role the banks do play in executing Beijing foreign exchange policy, where the money is going, and explore the possibility the banks are hoarding US dollars in less traditional ways.
Second, given the lengthy Chinese track record of creative data we must always be aware of the possibility Chinese data is less than accurate. For instance, despite treating Chinese bank financial statements as equivalent, Chinese banks state very clearly that they should not be considered equivalent to western banks even though they may appear as such. We must start with them as the source of truth and then if they are not a reliable source of truth demonstrate why and how, but for now we start with them as the source of truth.
China may produce output that generates a large trade surplus, however, there simply is no evidence that Chinese banks are materially increasing their non-RMB assets inside or outside of China.