My Blog

July 10, 2008

After I posted yesterday’s entry, one reader (RebelEconomist) asked the following question:

You often say China's money supply is growing fast because of its currency policy. Surely, if this was the case, base money growth would be strong. In fact, it appears to be very weak (growing about the same rate - 11%ish - as REAL not nominal GDP). It appears that sterilisation is completely successful. What indicator do you look at that makes you worry?

Since I have seen a lot of RebelEconomist’s writings and know him to be very strong on monetary issues, I hesitate to tangle with him on the subject, but it is a completely fair question and one I figured I should deal with formally in its own entry.

1.        China has a very high money base as a share of GDP.  I don’t remember the number (I will look it up some other time), but it materially exceeds that of most other countries, including most other large developing countries.  Increases, therefore, are from a much larger base, and any initial excess is likely to be multiplied.

2.        China sterilizes money creation by the issuance of central bank bills.  These are, however, extremely liquid and too close a substitute for money for them to reduce underlying liquidity in any significant way.  Corporations and banks hold them as cash and small investors eagerly buy up any on offer.  For these reason I see them mainly as a substitute of slightly more liquid form of money with a slightly less liquid form.

3.        I am very uncomfortable with the distinction between demand deposits and saving deposits, including long-term CDs, in the Chinese banking system.  I say this because a few months ago I purchased a 2-year CD at China Merchants Bank, which after one month I needed to convert (about one-quarter of it) into cash.  I expected a lot of difficulty and a penalty payment for breaking the CD (there had been no change in nominal rates, by the way), but it took less than 30 minutes, which is a dizzying speed for any transaction in a Chinese bank, and there was absolutely no penalty.  My only “cost” was I had to forgo the higher CD interest rate on the portion I cashed, although I still received interest on that portion at the demand deposit rate.  I have been told by friends that this is perfectly normal, so it seems to me that there is no real distinction, as far as I can see, between a demand deposit and any other kind of longer-term deposit except that you get a higher interest rate for the latter.

4.        In general for reasons I discuss below I tend to be very wary of the value of aggregates in determining money supply changes, especially in a country whose financial system is changing as rapidly as that of China, and more especially in a country that has a currency regime instead of a domestic monetary policy, so I prefer to use “gross” indicators like central bank liabilities.  This comes from many years of Latin American experience, especially in a country like Argentina with its “Convertibility Law”, a form of a currency board.  As money flooded into Argentina in the 1990s, it seemed to experience all the things we associate with a liquidity bubble far more than the growth in most of the higher-powered aggregates would have implied.  For this reason if PBoC reserve accumulation soars, I automatically believe money supply is soaring.

5.        Finally, the proof is in the pudding.  If I were to ask you what some of the consequences of excess money growth might be, you would probably tell me: rising inflation, low real interest rates, systematic overinvestment, highly speculative stock and real estate (and perhaps art) markets, rapid credit extension, and so on.  We are experiencing all of these things in China.

In general I tend to be a little Mundellian (at least according to my reading of Mundell) on the subject of money.  I don’t really think of liquidity as an aggregate, but rather as a quality.  All assets have greater or lesser degrees of “money-ness” and this can change over time.  What we usually call “money” is simply the asset with the most amount of money-ness, and although it is not always easy to decide what to include in our definitions of money, this reflects in part the problem of money as a quality rather than a quantity.

Even “money” can change its money-ness.  For example, not very long ago US dollars were a much more important part of China’s money base than they are today because they were – from what I have been told by old-timers – much more actively used as a medium of exchange and as a retainer of value.  People in China still have dollars (for example I do) but they are not used for everyday purposes at all, and they are costly to convert, so their impact on underlying liquidity in China is much lower.  When I lived in Haiti, however, dollars and gourdes circulated equally freely in shops and businesses at a constant exchange rate ($1 was then worth 5 gourdes – this was many years ago) and so it was only right to consider both of them equally as parts of Haiti’s high power money supply.

When an asset becomes more liquid and its value easier to ascertain at any given time, it becomes more money-like and it increases the underlying liquidity of the system.  A highly liquid and very widely traded and valued asset contributes more to the money base than one less so.  This was part of what Alexander Hamilton realized when he agitated for the creation of a single US national debt to replace the thousands of individual pieces of debt incurred by states and military during and after the Revolution.  His belief – well borne out by subsequent events – was that if these various obligations were assumed by the US central government, restructured into a single, large and fungible security, and given specific money functions (they could be used to pay taxes, as bank capital, etc.) they would essentially provide the US with a badly needed money base. It seems that he was right.

Perhaps a better example might be one cited by Charles Kindleberger.  After the Franco-Prussian War the victorious Prussians imposed on the French a 5 billion franc indemnity – then equal to about 25% of French GDP or 2-1/2 times the annual government budget – which the French were nonetheless able to raise fairly easily via a massive bond offering organized, I believe, by the Rothschild banks.  Kindleberger argues that the bond issue and subsequent huge transfer of (mostly) gold to Prussia represented a significant increase in European money supply because the bond itself fulfilled the function of money given its tremendous liquidity, high quality, and very diverse investor base.  Europe didn’t simply see a transfer of money from France to Prussia, in other words; it saw the addition of new asset that was a fairly close substitute for money and so a net increase in the “money supply”.  Not coincidently perhaps, immediately afterwards until the series of crashes and banking collapses that began in 1873, Europe and the world experienced a speculative frenzy that extended across the globe (in fairness the speculative frenzy began before the Franco-Prussian indemnity, but it reached its insane peak during and after).

When I lived in NY I used to discuss this often with the late Frank Fernandez, former Chief Economist at the SIA, and he often expressed his frustration at obtaining some reasonably reliable measure of global or domestic liquidity.  His conclusion, finally, was that there were no aggregates worth measuring.  In the end the best thing to measure was what he called the “shadow” of liquidity – such things as credit spreads, off-the-run Treasury spreads, bid-offer spreads for the most liquid assets, and so on.  I think I have absorbed his skepticism about the value of monetary aggregates as an accurate or even useful (in some contexts) measure of underlying money, especially in an unstable and rapidly changing financial system.

Comments (16) for "What is money growth in China?"

Michael, I am grateful that you gave my question your special attention! I think you must be right that deposit money does bid up prices to some degree, albeit less than base money. A lot of analysts remark that China's sterilisation is incomplete without actually saying what they mean. I asked my second question because I am wondering if China is quietly willing to tolerate some inflationary real appreciation of the renminbi as well as allowing the peg to crawl. It seems to me that inflation is less of a problem in China than in the developed countries, because in the developed countries "inflation" is the manifestation of a decline in their terms of trade. In other words, if developed country central banks did prevent cpi inflation, it would mean a fall in nominal wages. By RebelEconomist - 7/10/2008 1:25 AM

The consensus is that inflation scares the authorities, and of course the recedents are not good. The Party remember that hyperinflation in the 1940s probably did more than anything else to undermine support for the Nationalists, and more recently the inflations of 1985-87 and 1993-94 both led to social and political disturbances. I confess I am not totally sure why this has been the case, but certainbly most people believe that the authorities cannot let inflation be perceived as getting out of hand. By Michael Pettis - 7/10/2008 6:54 PM

Besides the liquidity of these CDs and other debt instruments contributing to money supply growth, don't bank holdings of these assets also increase the availability of credit by giving banks more reserves or do the frequent reserve requirement ratio increases cover (sterilize?) those bank reserve increases? By ross - 7/10/2008 7:41 PM

"The Party remember that hyperinflation in the 1940s probably did more than anything else to undermine support for the Nationalists, and more recently the inflations of 1985-87 and 1993-94 both led to social and political disturbances.I confess I am not totally sure why this has been the case, but certainbly most people believe that the authorities cannot let inflation be perceived as getting out of hand." The main reason is that Chinese people overall are net savers. Inflationa hurts them very badly. While in U.S. ordinary people seem to be mostly net borrowers, they benefit from inflation somehow. Things are changing in new century, more and more people take morgages to buy apartments and become net borrowers, however I heard most morgages in China are float-rated. By AChinese - 7/10/2008 10:58 PM

AChinese, good point and I am embarrassed I didn't bring it up. Of course unexpected increases in inflation will hurt savers, and Chinese households are high savers. By the way all Chinese mortgages are floating rate, and so increases in interest rates will reduce disposable income for mortgage holders. A friend at the PBoC told me this was a real concern of theirs. By Michael Pettis - 7/11/2008 5:26 PM

My reaction exactly Michael! But then you are American and I am British so we tend to forget about savers. Good point AChinese! By RebelEconomist - 7/11/2008 8:34 PM

Thanks for the interesting discussion. Of course, the other side of the monetary equation that is usually forgotten is velocity - which can only really be measured ex post facto. And it does change, particularly in places where there are other currencies and substitutes of money in use. And I much appreciate your discussion of the "quality" of money. How would you say this relates to velocity? Is it a measure of the same thing? GA By Gerg - 7/12/2008 2:52 AM

GA, I think you must be right; wealth will tend to be held in the form of money if it is likely to be used, so the effect that a form of money has on prices should depend on rapidly it turns over. A direct measure of velocity (eg from banks) is probably more useful than an estimate obtained from the equation of exchange (MV=PT). As you say, velocity will depend on what other liquid forms of wealth are available. I would not be surprised if the velocity of deposit money in China was lower than in, say, the US because it is used more for saving. Perhaps the fall in the stock market in China has accelerated the growth of deposits because small investors are reluctant to finance companies directly by buying shares and banks provide more company finance as the bond market is small? By RebelEconomist - 7/12/2008 10:41 PM

Rebel - yes, I agree, that the change from share ownership to deposits could accelerate velocity. As far as I can tell, this is the most difficult part to address in looking at emerging economies with unstable forms of "money". For example, a fall in the de facto backup currency (i.e. the dollar) can easily provoke a flight to the local currency, provoking a rise in velocity that may not be sustained in the long term. Then there will likely be some shifting between different types of money, with unpredictable effects, as all of the observable facts come with some delay... Frankly, I haven't seen a framework for analysis of this that seems sufficiently robust and simple. By Gerg - 7/13/2008 2:30 AM

Accelerate growth not velocity. Velocity would fall, because deposits held in place of shares would be relatively immobile, being held more for savings purposes than transactions. But perhaps you meant to say that. By RebelEconomist - 7/13/2008 5:29 AM

As important as it is I am nervous about trying to understand changes in velocity in a country like China, where so much change is taking place that has a direct impact on the way money is used. The banks are under intense pressure to improve the quality of service, and in the nearly seven years I have been here I have noticed huge changes. Cash is still king here, even for transactions measuring in the thousands of dollars (for example rent) and sometimes tens of thousands, but credit and debit cards (but not checks) are spreading very rapidly. Wire transfers seem quite common -- more than I remember them being in the US. A lot of Chinese kids use telephone payment cards to pay for a variety of things (my CD label is considering using them to allow downloads, already a common thing here). And of course we are still seeing large scale rural-urban migration, which changes the way people use money and banks. I am not sure how all these things affect velocity but I would imagine that it is fairly unstable, making the PBoC's job a little more difficult, as if they needed that. By Michael Pettis - 7/13/2008 1:47 PM

By the way, I said i would check the size of China's money base relativce to the economy. According to DB (http://www.dbresearch.de/servlet/reweb2.ReWEB?rwkey=u1562160), M2 was projected to be 164% of GDP this yer. That strikes me as high. Does anyone know what the equivalent numbers are for the US, Japan, Brazil, India, or other possibly comparable large countries? By Michael Pettis - 7/13/2008 1:52 PM

Just as I posted the above, the PBoC released June's M2 figure -- $6.49 trillion. Nominal GDP was $3.5 trillion last year and is expected to be $4.3 trillion this year. Taking the average, that puts M2 at 168% of GDP. By Michael Pettis - 7/13/2008 2:19 PM

Rebel - I actually did mean acelerate velocity, not growth, but it is arguable. My point is that any kind of massive change from one form of money to another (in your example, from the stock market to deposits) could cause a jump in velocity (even if only a blip). But the medium-term effect is ambiguous when measured overall (as opposed to just in the base currency). The classic case of a high velocity situation is when faith in the local currency disappears and everyone changes as quickly as possible into whatever foreign currency they can use most easily. What this means when there is a secular shift into deposits from other types of investment/wealth is ambiguous, but one could argue that since stock market wealth is not used to transact except on the stock market, that all else being equal a shift from stocks to savings accounts (which now are high-powered money) could indeed raise velocity. That said, I fully share Michael's caution about interpreting velocity in emerging markets like China. My point is, velocity is sort of the dark matter of the monetary relationship. I've heard this put stronger, such as the "magic" that explains what happens after you have all the other numbers, reducing the explanatory power of the monetary equation completely unless you can magically assume it is stable. And I guess my question is, is velocity the same thing as what Michael has described as the "quality" or "moneyness" of money? Or are they different? By Gerg - 7/13/2008 2:45 PM

Michael, I know the stats for the US and Japan: US M2 to GDP is 54%, Japan's is 146%. Gerg, as I said unclearly earlier, I do think velocity (actually the rate of turnover of the stock of money to accomplish the transactions covered) is the nearest thing to moneyness. By RebelEconomist - 7/13/2008 9:47 PM

It is also worth looking at the other side of the aggregate balance sheet - counterparts to money - can often tell an interesting story. I'm unsure of the level of the interbank gap, I'd imagine it would be high though don't know if it would be consistent in any direction. Would also agree strongly with AChinese's point. Inflationary spikes have been associated with political difficulty and subsequent tightening of economic policy. I have also seen it suggested that the spike in inflation in the early 90s was in-part politically encouraged in order to re-distribute some income to the poor. An interesting theory, just a theory, not sure if it's also relevant now: food price inflation benefits farmers, subsidies to the poor have been publicised (though I don't know of their absolute volume) and Hu Jintao seems quite the social reformer. By Alex - 7/14/2008 10:08 AM