My Blog

December 19, 2007

Friends of mine who try to keep track of monetary conditions by tracking central bank reserves – like Logan Wright and Brad Setser – have not always had an easy time of figuring out what is happening in China.  Since August, however, things have gotten more difficult.  At $20-25 billion a month, growth in reserves during the past four months has consistently come in lower than what one would have expected by adding together the current account surplus, FDI, and estimates of hot money inflows – year to date,. For example, monthly reserve growth has averaged nearly $40 billion.

Normally this would be a good thing.  The faster reserves grow, the more currency and central bank bills the PBoC has to create in order to fund this growth in reserves (the PBoC must buy the net inflow of foreign exchange in order to maintain the desired currency level – and while reserves are the asset side of the balance sheet, they are matched by RMB liabilities).  This forces an expansion in the country’s already excessively large money supply.  If reserves are growing more slowly, as they seem to be, this should mean that China’s money supply is also growing more slowly, which is an unambiguously good thing for the country at this point.

But it turns out that there are at least two major sets of transactions that are clouding the figures we have for growth in reserves and their impact is to reduce the headline foreign exchange reserve figure.  First, and most obviously, the PBoC has completed or soon will complete the transfer of $200 billion of its reserves to the CIC, where it will presumably be managed more aggressively.  This will reduce the headline foreign exchange reserve level by $200 billion. 

Will this represent a real $200 billion reduction in the amount of currency and central bank bills in circulation?  Yes, but its net impact on overall liquidity will be somewhat less.  The CIC was able to purchase the $200 billion in reserves with the RMB 1.55 trillion it received from the MoF, who in turn got the money by issuing long-term bonds to the market.  These bonds all or mostly sit on the balance sheet of the PBoC or a commercial bank intermediary, but as the PBoC gradually sells these bonds into the market, the net effect will be an exchange of highly liquid central bank bills for less liquid MoF bonds.  Although these bonds are still effectively part of the country’s money base, their net impact on domestic liquidity will be positive (i.e. they will reduce domestic liquidity, although not one-for-one)

The second set of transactions is much more confusing, at least to me.  Beginning in August of this year, as it has been raising minimum reserve requirements, the PBoC has also been forcing banks to hold at least part of their new required reserves in the form of dollars at the PBoC.  One simple way of thinking about this is that it is as if Chinese commercial banks have had to increase their required RMB reserves, and then have been forced to swap these new required reserves into dollars – so that they go from holding RMB in their PBoC required reserve accounts to holding US dollars in the same accounts.

What is the point of swapping RMB required reserves into dollars?  It seems to do two things.  First, holding dollars in China is a losing proposition because of expected RMB revaluation, and so by forcing commercial banks to use RMB reserves to “buy” dollar reserves, the PBoC is effectively transferring the future foreign exchange loss from its own balance sheet to that of the commercial banks (unless it has hedge agreements in place, in which case it has no impact on profits and losses).

Second, by forcing a transfer of dollars from the PBoC account into the commercial bank accounts, it reduces the headline foreign exchange reserve number and seems to imply that there is both less pressure on the currency and less money in domestic circulation.

But I don’t think either of these two implications is correct.  It seems to me that forcing banks instead of the PBoC to hold dollars says nothing about pressure on the currency.  I would argue that the total net inflows are exactly the same except for one thing – commercial banks have been asked to assume part of the PBoC’s normal functioning (i.e. to buy dollars so as to maintain the country’s foreign currency regime), and as they assume this function the resulting purchases of dollars are whisked off the PBoC balance sheet.  But nothing real has changed, except that now banks, instead of the PBoC, will be forced to assume the foreign exchange losses.

Second, what is the effect of this “swapping” of currencies on the domestic money supply?  None at all, it seems to me.  Banks will have exactly the same amount of loans outstanding as they did before the currency swap, and all the other monetary aggregates will be the same. This is because the required reserves held at the PBoC are effectively “dead” anyway, and redenominating their currency changes nothing real.

Headline reserves will be lower, of course, but Chinese money supply will be exactly the same as if the required reserves had never been swapped into dollars.  It is the existence of minimum required reserve rates that presumably has an impact on domestic monetary policy, not the denomination of those required reserves.  The PBoC could force banks every month to swap from RMB into dollars and back again, and the only real impact would be to change headline reserves every month.  Real money supply in China would not change at all.

If my analysis is correct, (and if any reader thinks it isn’t, please correct me) I think the policy to force banks to hold part of their required reserves in dollars does nothing to improve China’s liquidity position but makes the PBoC less transparent.  I am not sure how this helps.

By the way Logan Wright (“Paying for RRR Hikes in USD”, Stone & McCarthy, December 19, 2007) estimates that impact of this swapping might have reduced headline foreign exchange reserve growth by an average of just over $20 billion a month for the past four months – he assumes that the four hikes during this period took about 190 billion out of the system each time, and that this amount was “swapped” into dollars in the PBoC accounts of the various commercial banks.  With reported monthly foreign exchange reserve increases of $23.4 billion in August, $25.0 billion in September, and $21.4 billion in October, he may have explained rather niftily why monthly reserve growth fell off from the average of just over $44 billion in the first half of 2007. 

Comments (11) for "Currency swapping does nothi...

Michael -- I agree with your analysis of the monetary impact of forcing the banks to hold required reserves in dollars. Sterilization comes from the level of the required reserves, since as you note, funds on deposit in RMB at the PBoC are effectively. This effectively reduces the amount of sterilized intervention the PBoC does by forcing the banks to do the sterilization (Taking in RMB and buying $). It would be a bit different though if the banks were forced to hold fx as an asset, not as part of their required reserves. By bsetser - 12/19/2007 9:33 PM

"Banks will have exactly the same amount of loans outstanding as they did before the currency swap, and all the other monetary aggregates will be the same. This is because the required reserves held at the PBoC are effectively “dead” anyway, and redenominating their currency changes nothing real." Only if you can prove this. By fatbrick - 12/19/2007 10:25 PM

Dear Prof. Pettis, could you recommend a site to me where I could find some Chinese national accounts statistics which are at least mathematically consistent . I keep getting tired of reading news in the media, that all expenditure components of the Chinese GDP growing between 18-30% a year (investment >30%, consumption 18%(well, retail sales), trade sufficit close to 50%) while overall GDP growth is 10% and inflation is 3%? How come that nominal GDP growth is 13-14%, yet all of its components grow faster? By Gabor - 12/19/2007 10:38 PM

Correct me if I am wrong, as far as I know, the calculation of China GDP does not use the same method as that of US GDP. The way they collect data and compute is Not the expenditure method. I remember the method in China is more close to GDP = R + I + P + SA + W where R = rents I = interests P = profits SA = statistical adjustments (corporate income taxes, dividends, undistributed corporate profits) W = wages I cannot browse Chinese website now. You can take a look at Chinese Statistics agency's website. By fatbrick - 12/19/2007 11:08 PM

First of all, I think it is likely that the PBC swaps are hedged, because previous transactions in which the PBC forced banks to hold dollars were hedged against exchange rate risk. Second, forcing banks to hold dollars allows for something that can't otherwise be done, and that is to move the dollars "offshore" to Hong Kong. Once a bank is allowed to have dollars it can then transfer that money outside the "currency firewall" to Hong Kong and use it for overseas operations. Both dollar reserves held by SAFE and RMB deposits held by banks are "inside the firewall." To Gabor: My suspicion is that the difference has to do with depreciation of capital stock. By Twofish - 12/19/2007 11:12 PM

I heard an analyst said that the banks are getting a fixed exchange rate on $ (thus no danger of depreciation). So that removes that reason as a possible explanation for why has the PBOC told banks to put $ not yuan in reserve. Which leaves us with what feels like a very cynical (and poor) reason: This is a bad attempt to reduce the headline number for foreign exchange reserves. But why do so if it's so easy to see through (as apparently Logan Wright has). Or is the Chinese leadership just trying to spin the issue in the less-sophisticated media, which may not realize what's going on here? Or failing that, does anyone have another reason why they'd do this? I'm puzzled. By Chinawatcher - 12/20/2007 8:56 AM

Fatbrick, reseves held at the PBoC cannot be used for loans, so the fact that they are denominated in dollars instead of RMB will have no impact on their ability to lend. I can't prove that the total amount of loans are the same, but I think I can prove that the total amount of money available for lending is. Twofish and Chinawatcher, if the reserves are indeed hedged, I am as puzzled as anyone as to why they are doing it. I hope it is not just window dressing. Gabor, I use http://www.stats.gov.cn/english/ as my homepage. You can also get good stats on the PBoC website. Numbers are always a problem, and not just because of poor data collection. The fact that the economy is going through such major chnages makes it very difficult to maintain consistency of certain indices. By Michael Pettis - 12/20/2007 11:46 AM

Xu Jian (Kurt), one of my PKU students who now trades local markets for a foreign bank, wasn't able to add this comment and so asked me to do it for him: I think some other reasons may also explain the dollar reserve. First, China encourages the reform of fx system and wants to see a more flexible RMB rather than a one-way bet. So, the dollar reserve creates some buying interest in the market and push the rmb higher after every four hikes this year. I remembered the RMB is traded in a range during the last several reserve ratio hike announcement date and effective date. So is this time. As I talked to some traders, they fund dollar by both b/s $rmb swaps and buying spot. Second, China suspended onshore entities from trading offshore product at the end of 2005 to control speculation, but as far as I know, they don't want to prevent offshore deals forever as long as they are comfortable with the uncertainties and negative effects caused. Otherwise the new CFETS(China Foreign Exchange Trading System) won't integrate NDF function, even if this function doesn't go live at the moment. (by the way, all onshore $rmb spot/forwards/swaps are mandatorily done through the system). Due to dollar squeeze partially caused by dollar reserve, the onshore dollar curve is several hundreds pips higher than LIBOR and then drives on/offshore swap points quite close to each other rather than 1000-2000 pips gap in 1y NDF/DF some months ago. This action indirectly extinguishes speculators' activities between on- and offshore mkt. The third, PBOC doesn't need to issue as much bills as before to absorb the dollar and its fx reserve pressure eases... By Michael Pettis - 12/20/2007 12:17 PM

Let's see if I understand, Kurt. What you are saying is that by artificially forcing onshore USD rates higher by forcing banks to raise dollar funding, the PBoC is effectively pushing the onshore dollar forward rate lower (since onshore forwards are priced against interest differentials, unlike the offshore NDF market). This makes it less profitable to buy dollars forward on the offshore NDF market (where of course they trade at a discount to spot, since the dollar is expected to depreciate against the RMB) and sell dollars forward onshore. If I have interpreted you correctly, then bravo! The PBoC, by forcing banks to hold dollars in their reserves, is creating an artifical demand for dollar funding that drives USD interest rates higher onshore, making it more expensive for domestic speculators to take advantage of the arbitrage. This may reduce some of the hot money inflows. Am I right, Kurt? If so, very clever. By Michael Pettis - 12/20/2007 12:26 PM

But then don't they eventually have to unwind these reserve holdings and just reverse the process, causing even more problems? Or is the thinking that the dollar reserves are being held long term and that by the time they get around to "unfreezing" the dollars they'll be beyond the current currency issues? By Chinawatcher - 12/20/2007 3:52 PM

By Chinawatcher - Thu, 20 Dec 2007 23:52:16 GMT As Kurt said, when the real economy conditions change, i.e. USD appreciates again, they can safely reverse the process without a major impact, when "they are comfortable with the uncertainties and negative effects caused". As I understand, with a huge reserve of USD, the 2005's decision to stop offshore bet, and the closed capital market, virtually PBOC acts as a cnetral bank to issue USD in China. The interest rate in USD is higher so people are willing to hold USD deposit in China. This might explain why all major banks offer USD investment products with high interest rate in China these years. By fatbrick - 12/21/2007 1:09 AM