Yuan fix hits new low, capital flight intensifies

Tracking overnight markets the PBOC yuan fix has hit a new daily low today:

Interesting given the US dollar fell last night. The devaluation is, if anything, accelerating.

And so is capital flight, from Goldman:

China capital flows update—sources how cross-border RMB flow might mask outflow pressures

• We have updated our estimates of sources of China’s capital n outflows. Our analysis suggests net capital outflows at $123bn in Q1 (vs. $504bn in Q3-Q4 combined last year).
• Of the Q1 net outflows, about 70% was due to Chinese residents’ accumulation of foreign assets; 40% to repayment of FX liabilities; and -10% to foreigners’ demand for RMB assets (i.e., foreigners were a source of net inflows in Q1). This composition is broadly similar to our earlier estimates for 2015 H2.
• Separately, we flag a large $170bn net RMB flow from onshore to offshore since last October, which has helped reduce FX reserve drawdown and put downward pressure on CNH forward points. This flow cannot be readily explained by marketbased factors in our view, and did not seem to result in an increase in foreigners’ CNH holdings. We think it might have masked the true FX outflow pressure in China, on the order of some $20bn (or 50%) per month in recent months.
• Going forward, we think it will be important to also track cross-border RMB movement to get a fuller picture on China’s underlying flow situation.

While according to the BOP the pace of capital outflows has slowed in Q1, it might not have in fact slowed by as much as the data suggest. We have in the past discussed various caveats to interpreting official flow and reserve data, and in the following we add one more, in light of a large unusual cross-border RMB flow in recent months that we believe could have masked the true outflow pressure in China.

A $170bn flow of RMB to offshore…

Specifically, since October last year we have seen a large net flow of RMB from onshore to offshore, primarily due to trade settlement in RMB (i.e., Chinese importers pay for the imports in RMB). This totaled $170bn through May or about $20bn per month on average (Exhibit 4). This flow has helped lessen the overall outflow pressure faced by China because it means that importers did not have to buy as much FX to pay for imports (since they just used RMB). This also helps explain in our view the general decline in CNH forward points (or equivalently, CNH interest rates) in the last few months (Exhibit 5), despite market perception of large-scale CNH smoothing operations by state-related entities (more on this below).

Compared to previous actions, this is somewhat unusual. In the past, net crossborder flow of RMB had typically been driven by offshore RMB sentiment, e.g., when offshore RMB sentiment is strong, CNH tends to be more expensive than CNY ($/CNH is below $/CNY), naturally driving a net flow of RMB from onshore to offshore (e.g., for trade settlement) to satisfy high RMB demand; and vice versa.

However, especially since the August 2015 RMB reform, offshore RMB has been generally weak. While the CNH-CNY gap has narrowed in the last few months, CNH has still been usually cheaper than CNY ($/CNH above $/CNY). Therefore, the typical market-driven relationship would have suggested a net flow of RMB from offshore to onshore instead. Indeed, in a stark contrast, the relationship is in total reverse since October last year—the cheaper the CNH (vs. CNY), the greater the net flow of RMB from onshore to offshore. This is more consistent with a supply-push pattern (an exogenous push of RMB from onshore to offshore, which causes CNH to trade cheaper), rather than a market driven demand-pull relationship.

In short, we cannot point to any obvious market forces that could explain the RMB flow in the last several months; non-commercially driven factors seem to be a more likely explanation, in our view.

… that does not seem to result in any increase in foreigners’ CNH holdings

Another interesting observation is that this large amount of net RMB flow to offshore does not seem to show up in foreigners’ holdings of CNH assets. In general, if the RMB is received by foreign non-banks, that would likely end up as CNH deposits; and if it is received by foreign banks, that would show up as an increase in banks’ holdings of CNH assets.However, CNH deposits in Hong Kong and Taiwan, two key CNH centers, have been on a decline in the last several months (Exhibit 7); and Hong Kong banks’ spot position of “other currencies” has also been falling (Exhibit 8).

More broadly, overseas entities’ holdings of onshore RMB deposits (which include placement of CNH by offshore banks to onshore banks) have as recently, sharply deviated from the hitherto synchronized pattern with the cumulative net flow of RMB from onshore to offshore, and have been even surpassed by the latter in absolute level (Exhibit 9).

Given the discussion above, it is possible that the actual underlying FX flow situation (i.e., FX/RMB demand by Chinese corporates/households and foreigners) has been less encouraging than even the SAFE data on onshore FX settlement imply10. (e.g., according to that data alone, FX outflow was about $2bn in May.), but there was also $24bn in RMB flow to offshore during the month—if we assume that that flow was not market-driven and that it was not subsequently held by foreigners, then the underlying FX outflow could instead be $26bn in May. In the eight months since last October, this approach would have suggested a total net FX outflow of about $500bn, or 50% above the $330bn implied by SAFE’s onshore FX settlement data.

Going forward, we believe also tracking the data on cross-border RMB flow (released monthly by SAFE) will be important to coming to a more comprehensive view on the underlying flow picture. For the time being, we will be incorporating this into our measure of net FX flow (Exhibit 10 shows this modified version