Brad Setser at the Council of Foreign Relations has a good piece on the Chinese FX position with an interesting point about the state of Chinese bank FX holdings. He makes the very interesting point that Chinese depository corporations foreign assets have continued rising pretty much on trend for quite some time, but after August 11, foreign liabilities of banks have plunged. He posits that this is a good thing, indicative of financial strength via rapid increase in net FX holdings, and that the PBOC has higher level of implied FX reserves than is understood.
I think there is another much more likely explanation that is supported by the data that leads to a different conclusion.
Before we even dive into the data, think about the point that Chinese bank foreign assets have risen effectively on trend (an important point) but foreign liabilities have dropped significantly. On the face of it, this should strike you as very odd. The primary input for a bank is either deposits or liabilities that they then use to lend or purchase a fixed income asset. If a bank has significant drop in its input, how does it maintain trend growth of its output? Put another way, where are Chinese banks getting the foreign currency (deposits or liabilities) they use to increase foreign assets?
Let me reframe this away from banking. What if Starbucks reported that coffee drink sales had doubled but they also reported a 50% fall in bulk coffee purchasing? Would seem on its face a little odd. Had prices changed significantly? Had they changed their formulas? What was happening to cause sales of coffee and purchases of coffee to go strongly in the opposite directions? That is effectively what is happening here.
So this leads us to dive into the data. How are Chinese banks funding foreign asset purchases while reducing foreign liabilities? Where is the foreign currency coming from?
The rapid drop in foreign liabilities is likely disguising capital outflows and hiding debt. I know of Chinese and major MNCs that are effectively being blocked from engaging in FX transactions but allowed to conduct a variation on this theme. Here is how this happens. A company wants to move money out of China but is refused the FX so is forced to keep RMB in China. A bank, typically a major bank, offers to arrange the transaction for them like this. The client deposits money at the bank offering the cash as collateral. The bank arranges for a swap with an offshore entity to then lend USD/EUR/JPY whatever the client wants in the jurisdiction, backed by the secured cash. There is no explicit movement of capital between China and other jurisdictions and there is no foreign currency liability.
It must be noted that while we cannot say with perfect certainty this is what is happening, all evidence supports this hypothesis. Besides the anecdotal evidence let me give you some supporting data. Bank of China and ICBC (PDFs) in their 2016 annual reports give evidence of this behavior. BoC’s and ICBC’s notional amount of FX swaps grew by $125 billion and $69 billion. In other words, the amount of money that they have worked to provide swaps for, in these two banks alone, is up almost $200 billion in 2016.
Market data supports this move to supporting outflows via the swap market. In January 2015, turnover in FX swaps was about a third of the spot market. In between August and October 2015, the FX swap and spot market equalized (notice the timing) and now the swap market is about one third larger than the spot market. Since May 2015, FX spot market turnover is up a pedestrian 15%, but FX swap turnover in China is up 73%.
But wait, there’s more! FX spot market transactions between banks and their clients from May 2015 to 2017 is down 6% while interbank FX swap volume is up 82% during this same time. Now the interbank FX swap market is 271% larger than the FX spot market for bank clients. Then we see that Chinese banks are significant net buyers from customers of FX in the spot market. Taken together this implies that Chinese banks are soaking up hard currency into China and arranging for outflows via FX swaps that do not actually facilitate currency flows from China to the rest of the world.
It is worthy to note that while many people believe the RMB has gone global, most central banks hold minimal if any RMB. What they have are currency swap agreements that allow them to access RMB when needed and the PBOC to access foreign currency when needed. Given that bankers inside and outside of China treat BoC as effectively a branch of the Chinese Ministry of Finance, it is likely BoC engaging in various types of swaps agreements to give it overseas hard currency funding sources that keep its primarily liabilities in RMB.
There are a few final points of note. First, if Chinese banks moved rapidly out of actual foreign currency liabilities and into swaps to fund overseas asset purchases, this would explain the trend growth in bank foreign assets but the drop in liabilities. Swaps are not accounted for based up the notional liability amount but on a “fair value basis”. If the banks engage in currency swaps and then use the currency to fund foreign asset purchases, this serves to effectively undercount the liability by carrying it at fair value and double counting the asset at 1+fair carrying value.
Second, it is important to note that depending on exactly who is holding these swaps and how balanced the book is, this implies that the FX has not fallen at all if there is sudden movement in the RMB. These are simply implied liabilities. For instance, BoC is carrying FX swaps equal to a notional value of 5.36 trillion RMB or nearly $800 billion USD but they carry these on their books as liabilities equal to only 87 billion RMB or $13 billion USD. The accounting value is equal to 1.6% of the notional value. While on the face of it this appears relatively standard accounting value liabilities, it is important to note this underlying issue.
Third, if the PBOC needed to access Chinese bank assets, their net asset position is being overstated. The foreign currency can fund loans for foreign asset purchases that are recorded on group balance sheets as loans to customers but record only a fraction of the liability used to raise the foreign currency overstating the net asset position. It would also appear to overstate the liquidity of such assets if the PBOC ever needed to coordinate such actions.
I hope this is clear as these are some more technical issues. However, I think it is fair to say that this is much more likely scenario that does not lead to such a rosy outcome.