RMB update (ii) - Implied basket weight

According to Oriental Daily HK, PBC continues its effort in making the new basket mechanism more transparent

  • Transaction volume will be announced annually, once in every May. (this is actually a step backward from the daily disclosure prior July 21st. But I presume it is part of package of the basket peg.)
  • Other statistics will also be released regularly: Current Account Balance (semi-annually - did I translate this correctly?), Foreign Debt Stats (Quarterly), Cross exchange of foreign currencies vs USD in China (monthly on 20th), inter-bank forex volume (annually), etc.

Stephen Jen of Morgan Stanley had some initial analysis of the RMB data

  1. "USD/CNY has been extraordinarily stable. Since July 21, 2005, USD/CNY has declined by a cumulative one-sixth of a percent, or 0.15% — half the daily allowable range.
  2. Implied USD weighting very high. An alternative way to justify the stability of USD/CNY is to calculate the implied weighting on various currencies. Based on the actual movement in USD/CNY, it is as if the USD weighting is some 85% or so of the basket. "
  3. Transition period. It is possible the PBoC is only gradually allowing the basket index to be revealed so USD/CNY remains stable in the initial period of the BBC. "
  4. There is no mechanical link between the basket weighting and the currency composition of reserves. ...although China has a basket reference scheme, the USD will remain the sole intervention currency.
  5. The PBoCÂ’s holdings of JPY in its reserves are likely to be substantially below 18%.
  6. I still do not grasp why Governor Zhou announced the content of the basket. The desire to guide domestic investors to learn to hedge aside, I am still not convinced it was a good idea to announce the content of the basket. If it was such a good idea, why, then, was it not announced on July 21, 2005?"

The "Implied USD weighting" means the result of multi-variable dimensional regression. Since I do not have access to the international rate at the exact minute, I could not do an accurate analysis like Jen did. But this is something we could all do with an Excel spreadsheet (without using regression tools).

  1. Go to PBC's site (in Chinese, the column headings are Date/USD/EUR/JPY/HKD) copy and paste the data into your Excel.
  2. Define weight factors: wu,we,wj,wh for USD, EUR, JPY, HKG(which is aligned with USD). Let's ignore other currencies such as KRW for simplicity (also because we lack the data)
  3. Calculate for each day this number: RMB(date)=usd/8.11*wu+eur/10*we+jpy/7.3*wj+hkd/1.05*hkd (or instead of these values, divide by the previous day's rate)
  4. Now try different weight factor values for wu/we/wj/wh, until you find the values for all these days are closest to 1.00, you have found the approximate value of the 'implied weight'
  5. It helps if you create a cell to calculate sum((1-RMB)^2) over all dates, then if this number is minimum whether you change the weight factor up or down, then you have found the "implied weight" for that currency
  6. (I am happy to email you the spreadsheet if you are interested - don't know how to upload table/chart in blogger)

I found USD weighting to be even higher than the "85%" number which Jen mentioned. It is probably because SGD, KRW/etc are closely linked to USD (mathematically it is called linear dependency). In fact, it is hard to calculated USD's rate this way because there might have been some intervention and that the gaps for USD is different from (smaller than) those of EUR and JPY. Other findings

  • HKD seems to be not in the basket at all (the value of RMB does not change if you change the number)
  • JPY implied weighting likely to be around 10%, EUR perhaps only slightly larger
  • The fact that RMB is still so highly correlated to USD is still puzzling. Maybe PBC is smoothing out the transition, by adjusting the USD basket weight slowly from 100% down to the the target weight of 50% or 43%. In other words, in July, USD might have still been the sole content inside the basket, or RMB continued to peg to USD alone for a few more days, until the peg is slowly loosened. If Jen's implied weight at 85% is correct (the average over the period), USD weight might have decreased from 100% down to around 70-80% now, via gradual adjustments

My guesses as to why Zhou decided to reveal the basket content are:

  1. to make sure his people (and his superior) cannot mess up with the system (e.g., by over-riding the mechanical rules)
  2. he is not too worried about reverse engineering the content because there are other uncertainties such as the different gaps for different currencies, the drivers for the weights are not disclosed, and also that the weight mix will be revised sooner that it was figured out (and as I mentioned previously, speculators need ultra-accuracy to profit from their highly leveraged bets)
  3. rebuff at the "black box" allegation

Updates (Aug 23)

  • Johnson Electric Holding, the market leader in micro-motor (powers 50% of the car windows, windshield vipers in the world), said, their cost will rise by 0.2% after the 2.1% RMB reval (i.e. only labor, rent and some G&A, while the raw materials, copper and steel are largely unaffected)
  • Yu Yuen Industries (largest suppliers to shoe brands such as Nike) said the impact is 0.5%, labor content for sports shoes is higher than assembling micro-motors
  • These are exactly as predicted in my earlier post, I am sure the productivity improvement per year for these companies is faster than the rise in cost due to RMB reval. In fact, isn't the reval an effect of improvement in productivity?


Sun Bin said...

could you send me your email address? or post as comment here (i will delete it as soon as i got it).

You can also delete it yourself right after posting (I think), because I have set up the comment function to copy to my email account.

Myrick said...

Interesting stuff. I'd love to take a look at your spreadsheet as well (asiapundit-AT-gmail.com). Also of interest is implications for treasuries if the weighting of USD is to be reduced. China's gradualism could give it some leverage in other matters, as the Treasury (and one assumes the administration) does realise what a re-weighting could mean.

Sun Bin said...


I emailed you the spreadsheet. But we probably won't be able to interpret the data until there is major change in the USD/JPY or USD/EUR rate (i.e. over 0.3%), so that the gap would not mask the underlying drift.
Regarding implications on treasuries, I tend to agree with Jen's comment. Theoretically there is no correlation between the basket and the underlying asset. But there is pragmatic reason to overweight in USD.

We can use the comment field here to discuss.

Brad, will email you once I got your email address.

Sun Bin said...

Chi-hung Kwan at REITI (Japan) has a great essay (written a year before 7/2004) about why RMB reval is good for China.

Sun Bin said...

via New Economist, BOJ has a good paper comparing China's and Japan's currency reform.

Karen said...

could you pls send me the spreadsheet to calculate Implied basket weight for RMB? Very curious to know. Thanks,